Abstract
Machine and deep learning are used worldwide. Machine Learning (ML) and Deep Learning (DL) are playing an increasingly important role in the healthcare sector, particularly when combined with big data analytics. Some of the ways that ML and DL are being used in healthcare include Predictive Analytics, Medical Image Analysis, Drug Discovery, Personalized Medicine, and Electronic Health Records (EHR) Analysis. It has become one of the advanced and popular tool for computer science domain.’ The advancement of ML and DL for various fields has opened new avenues for research and development. It could revolutionize prediction and decision-making capabilities. Due to increased awareness about the ML and DL in the healthcare, it has become one of the vital approaches for the sector. High-volume of unstructured, and complex medical imaging data from health monitoring devices, gadgets, sensors, etc. Is the biggest trouble for healthcare sector. The current study uses analysis to examine research trends in adoption of machine learning and deep learning approaches in the healthcare sector. The WoS database for SCI/SCI-E/ESCI journals are used as the datasets for the comprehensive analysis. Apart from these various search strategy are utilised for the requisite scientific analysis of the extracted research documents. Bibliometrics R statistical analysis is performed for year-wise, nation-wise, affiliation-wise, research area, sources, documents, and author based analysis. VOS viewer software is used to create author, source, country, institution, global cooperation, citation, co-citation, and trending term co-occurrence networks. ML and DL, combined with big data analytics, have the potential to revolutionize healthcare by improving patient outcomes, reducing costs, and accelerating the development of new treatments, so the current study will help academics, researchers, decision-makers, and healthcare professionals understand and direct research.
Similar content being viewed by others
Avoid common mistakes on your manuscript.
1 Introduction
Nowadays, machine learning (ML), and deep learning (DL), are leading topics and centres of interest in the industry, academia, and popular culture all over the world [1, 1]. Machine learning has emerged as a popular field of study and cutting-edge tool in recent years. Multiple areas of prediction and decision-making benefit from its suite of algorithms and statistical techniques. The fact that it can use a variety of learning approaches means it can process large amounts of unstructured and complex data in ways that lead to revolutionary shifts in perspective based on the data experience.
Different machine learning-based model has been applied in two ways: (1) Supervised learning, and (2) Unsupervised learning. The data are input into the supervised learning-based model, and then the model is used to perform classification, regression, and prediction on the labelled dataset [3,4,5]. In contrast, an unsupervised learning-based model is applied to unlabeled datasets to perform tasks such as clustering, dimensionality reduction of a dataset, detection, etc. Also, the concepts of machine learning are being applied in various industries; it is largely responsible for the revolutionary changes occurring in the healthcare industry. The majority of the time, it is utilized for the prediction of early-stage disease, the identification of disease, and the monitoring of the patient's status based on clinical datasets. Detecting automatic object features in medical images can also be accomplished by the use of machine learning techniques to datasets related to medical imaging. To manage picture collections and larger datasets, the deep neural network-based technique has emerged as the model of choice among the various machine learning models [6, 7]. Because it uses the three-layer perceptron for analyzing the dataset. Deep Neural Network-based approaches are a great solution in larger-scale datasets with different structures in fields of data mining, Image processing, Natural Language Processing, Expert Systems and Prediction, etc. [8, 9].
The subfield of machine learning known as "deep learning" employs iterative methods for learning new features. It offers several neural network algorithm variants that can learn features in a variety of ways [10,11,12,13]. It is used on a huge scale and processed via numerous layers of filters. As a result of its use of multi-layer filtering techniques, it can deliver increasingly precise results compared to conventional machine learning while efficiently organizing massive datasets.
Healthcare is the fastest expanding sector, drawing in more scientists, professors, and medical experts who are all eager to make important contributions to the discipline and the healthcare system as a whole. Hospitals, medical equipment, clinical trials, telemedicine, medical tourism, health insurance, and other related services are all examples of healthcare facilities. It presents a tremendous chance to enhance healthcare facilities with the use of cutting-edge technologies. Therefore, the development of more sophisticated technology has allowed for massive amounts of data and proven to be a massive transformation in healthcare fields, in terms of providing easy access to the best diagnostic tools, the most cutting-edge treatments, and a variety of minimally invasive procedures that result in less pain and quicker healing. Drug development, tailored treatment, robotic surgery, illness monitoring, etc. all rely heavily on healthcare in various capacities. A field's research dynamics and behaviours can be gleaned from scientific articles using bibliometric analysis, making it the most reliable tool for spotting patterns in the research landscape. Mathematical and statistical methods are used to research articles, books, and other forms of scholarly communication, and the results are used to analyze patterns in the scientific publishing world. Systematics examination of published sources is commonly used to gauge advancements in science across a range of disciplines using a variety of criteria.
In this research, bibliometric analysis has been applied to analyze the most significant and crucial research pattern in machine learning (ML), deep learning (DL), and the healthcare field, as well as the most relevant research areas of machine learning (ML) and deep learning (DL) application applying in the healthcare field. Also, explored the most related research areas of healthcare like disease detection, diagnosis, and prediction [14,15,16]. For this, Thomson Reuters’ Web of Science (WoS) (Clarivate analytics, 2020) database has been used to collect the bibliometric information in topic-wise and title-wise categories and “Machine learning”, “Deep Learning” and “Healthcare OR health*” used as query keywords. The web of science (WoS) database is a well-structured and differently indexed database, that selected only the most relevant top publication, it includes many significant scientific articles. Out of all relevant significant-top publication document types, only article, review, and early access articles have been selected for descriptive analysis. Another aspect of the bibliometric study is the following: number of published papers from 2010–2021; the number of authors; the number of papers per author and contribution; authors' collaboration; country-wise numbers of publications: country-wise paper impact factors; year-wise paper publication; top-most institution etc. Finally, we elected top categories to show the research dynamic, trends, hot topics, and research directions and variations on this related type of strategy have been successfully applied by the recent literature [14, 17, 18].
Therefore, the purpose of this study is to give depth insight and a clear understanding of the research pattern of machine learning and deep learning application in the healthcare sector and compare the application of both techniques in the advancement of healthcare treatment. Hence, this study will be more beneficial for academia, and researchers’ health professionals to determine the relevant area of research in Machine Learning, Deep Learning, and Healthcare that has been broadly focused on along with the gaps that should be addressed. The paper is organized as (1) the “Introduction” section deals with the introduction of the topic. (2) “Methodology" section discusses the proposed methodology. (3) “Results and discussion” section provides the results and follows them with discussions. The “Summary” section deals with the summary of the whole work. (4) “Conclusion and prospect” section concludes and future scope of this work. (Fig. 1).
Workflow diagrams
2 Material and Methods
2.1 Data Collection
For this study, all bibliometric information has been collected using the text keyword search strategies in topical and title wise search from the Web of Science (WoS) database in machine learning and deep learning application in healthcare. The bibliometric information was searched only from Science Index-Expanded (SCI-Expanded) journal categories and retrieved complete data in plain text and BibTex data format, at global respect published literature during timespan 2010–2021(data accessed:31 august 2021). Then, filter the bibliometric information from Research articles, review articles, and early access articles from literature publication categories. Further, the bibliometric analysis method and network visualization methods were used to show the deep research trends and different levels of collaboration among scholars. In this context, the latest version of RStudio4.1.0 software was installed and then established the bibliometrics package [19] with R environment to perform descriptive bibliometric analysis and map the data. Then, the VOS viewer [20] open-source graphical user interface was used to demonstrate the show of different collaboration networks and bibliometric coupling based on retrieved bibliometric data.
2.2 Finding From Bibliometric Data Analysis
Table 1 shows the data retrieved from the database with searched keywords such as: “machine learning”, “deep learning”, “healthcare”, “machine learning” AND “healthcare”, and “deep learning” AND “healthcare” in topical and title-wise categories. This table data shows that, in both the categories such title wise and topical wise categories, the highest literature published with the keyword healthcare (1,67,326), followed by machine learning (98,169) and deep learning (51,559). Likewise, in the application field respect, machine learning techniques have largely contributed to the healthcare domain more than deep learning techniques, which is shown by the literature data.
Figure 2 represented the bibliometric information between the numbers of publications vs categories. It also shows that topic-wise literature is more than title-wise, which is shown by the blue and orange colours.
Statistics Of Categories Wise Data Pattern
Table 2 illustrates the total literature published research documents in various types, then three types of documents i.e., research article, review, and early access article were selected in both topical and title-wise categories. In a topic-wise published scientific article, the highest numbers of published scientific articles with keywords healthcare (article:1,28,247, review:22,071, early access article:3,540), followed by machine learning (articles:86,292, review:5,794, early access:2,854), deep learning (articles:46,217, review:2,412, early access:2,130), machine learning in healthcare (articles:1,625, review:316), early access:98) and deep learning in healthcare articles (750), review (149), early access (74).
Whereas, in title-wise published literature, the maximum number of research published with the keyword machine learning (article:22,935; review article:1,449; early access:22,935) followed by healthcare (articles:21,545; review article:2,730; early access:694), deep learning (articles:16,055; review article:730; early access:788), also machine learning in healthcare (articles:74; review article:11; early access:10) and deep learning in healthcare (articles:42; review article:5; early access:9). Figure 3 presents the statistical diagram, which demonstrates the published literature under different categories.
Total literature publication under different categories type
2.3 Data Searching Strategies
Then, data searching strategies were used to systematically search bibliometric information in topical and title categories. For performing these tasks single and combinational text keywords were used “machine learning”,” deep learning”, “healthcare”, “machine learning” AND “healthcare” and “deep learning” AND “healthcare”. The task was executed to gather statistical facts, research patterns, and trends regarding machine learning, healthcare, also, the uses of machine learning methods related to the healthcare domain based on bibliometric information.
Further, Inclusion parameters were set to include only peer-review literature from SCI-Extended journals available in the WoS database and described the machine learning, healthcare research area. From different categories of published articles in the WoS database, only three publication categories like research article, review article and early access article were included for this study in global respect. As Exclusion criteria, Social Sciences Citation Index (SSCI) and Arts & Humanities Citation Index (A&HCI), and literature published before the year 2010 were excluded. all country-specific selections were discarded from the parameters. Furthermore, an article published in Proceedings Paper, Book Chapter, editorial material, letter, correction, re-print, etc. were excluded.
Furthermore, to understand the statistical research pattern in the application of machine learning and deep learning related to healthcare, bibliometric data were retrieved from the WoS database in a bibliography file (.bib) and plain text format (.txt). Both files included information like (i) Authors, (ii) Abstract, (iii) Addresses, (iv) ISSNs/ISBNs, (v) IDS numbers, (vi) Funding Information, (vii) PubMed IDs (viii) Titles (ix)Cited References, (x) Times Cited (xi) Cited Reference Counts, (xii) Language (xiii) Sources, (xiv) Documents Type, (xv) Keywords, (xvi) Source Abbreviations, (xvii) Author Identifiers, (xviii) Article Information, (xix) Publisher Information, (xx) Research Areas, (xxi) Usage Counts, and (xxii) Highly Cited.
2.4 Methods Used
The bibliometric analysis and Visualization method has gained immense popularity in research at present time. Because it provides the right direction to the research community, academia, and health professional based on depth knowledge of scientific publication patterns. This study included the top 20 latest publication patterns, domains, journals, authors, countries, organization research activities, and recent trending topics.
For this, the Bibliometrix R package [19] is used for mathematical and statistical calculation of publication frequency, percentage, and citations of each author, journal, country, etc. A global collaboration map and other visualizations were done by open-source (corresponding information, author cooperation) VOS viewer [20] graphical user interface-based software, the tool was used for mapping of literature and analysis of author collaboration, country collaboration, etc. network visualization.
3 Result and Discussion
Bibliometric analysis has been conducted on the retrieved bibliometric data from all fields i.e., topic, title, abstract, affiliation, journal, etc. categorized by WoS core collection database in the research area of machine learning and deep learning application in the healthcare domain. To understand the various statistics of research trends in this research area, the latest relevant 1000 records have been selected among all available records from document-type research articles, review articles, and early access articles in SCI-Extended journals among globally available records. In this, the total number of research articles is 808, the total number of review articles is 120, and the total number of early access research articles is 45 in machine learning in healthcare, likewise, the total number of research articles is 806, the total numbers of review articles 106, and total numbers of early access research articles 71 in use of deep learning in the healthcare research field at worldwide level. At the Indian level, the total number of research articles is 213, the total number of review articles is 55, and the total number of early access research articles is 37, in the area of machine learning application in healthcare and the area of deep learning in healthcare, total numbers of research articles, total numbers of review articles and total numbers of early access articles are 157, 24, 38. Furthermore, the total number of research articles published in the field of computer science is 1057 articles, 611 articles at the worldwide level, and 137 articles, 104 articles in the research area of machine learning and deep learning in healthcare respectively.
3.1 Year-Wise Published Literature
Table 3 illustrates the year-wise scientific literature production at the worldwide level under peer-reviewed journals. So, the use of machine learning techniques in the healthcare domain has started in 2010, and the growth of implementation was very less till 2014 and after 2015 continuously increased till now. whereas the use of deep learning in the healthcare domain has started in 2015, and rapidly increased up to the highest peak in the machine learning field.
Hence, the higher scientific production rate shows the higher uses of deep learning techniques in the healthcare domain, compared to machine learning. Currently, deep learning technology is a more heavily used technology to handle more complex and unstructured data. The average year from publication is 1.5 and 1.2 around machine learning and deep learning used in the healthcare domain.
Figure 4 exhibits the year-wise publication or literature on the application of machine learning and deep learning in the healthcare domain. In this diagram, blue and orange colour sticks represent the area of research i.e., machine learning, and deep learning. So according to the graph, the research around machine learning application in the healthcare domain has been studied from the year 2010, so from years 2010 to 2021 the uses of machine learning techniques has continuously increased, similarly, research in the area of deep learning application in healthcare has started from 2015, so implementation of deep learning techniques in healthcare domains has rapidly increased. The diagram also reveals that deep learning techniques are highly used in healthcare to provide a better quality of medical treatment.
Year-wise publication pattern
3.2 Top 20 Most Relevant Sources of Publication
Table 4 demonstrated the top 20 most relevant sources of publication in machine learning and deep learning application in the healthcare domain. hence top 5 journals' literature publications in IEEE Access(articles:52), Sensors (articles:29), Journal of healthcare engineering(articles:22), Journal of biomedical informatics(articles:20), PLOS One(articles:20) in machine learning application in the healthcare domain. similarly, for deep learning applications in healthcare, the topmost 5 publication sources are IEEE Access(articles:78), journal of healthcare engineering(articles:35), sensors(articles:27), IEEE Journal of biomedical and health informatics(articles:22), scientific reports(articles:20).
In both areas of research, IEEE Access and the journal of healthcare engineering produces a higher rate of literature on the application of deep learning in healthcare.
3.3 Year-Wise Production of the Source of Publication
The year-wise scientific production of the source of publication has shown in the figure. It revealed the frequency of published articles in different journals. Hence, IEEE Access journal production started slowly from 2010, and after the year 2015, it rises continuously in literature production and has maximum in the year 2020 in the application area of machine learning in healthcare. The Journal of healthcare engineering articles’ production rate was slow from 2010–2017, and after 2017 its publishing rate increased and in the year 2021, it has a maximum of 10 articles. Similarly, in the present scenario total publication of the source journals are the journal of the American medical informatics(articles:5), PLOS One(articles:6), BMC medical informatics and decision making(articles:7), an international journal of medical informatics (articles:3) (fig. 5)
Year-wise production of a source of publication in machine learning in healthcare
Figure 6 demonstrated the year-wise publication rate of journals/sources in the application of deep learning in healthcare. The production of articles by top journals is IEEE Access(articles:26), journal of healthcare engineering(articles:26), journal of biomedical informatics(articles:2), IEEE journal of biomedical and health informatics(articles:10), scientific reports(articles:7), etc. among all top 10 journals IEEE Access production rate is sharply raised after the year 2017–2019 and maintained till the year 2021. Also, the journal of healthcare engineering started raising from the year 2017–2020 and after 2020 reached its highest peak. And rest of the journal production ranging (articles:2–15) from the year 2017–2021.
Year-wise production of the source of publication in Deep learning in healthcare
3.4 Top 20 Most Globally Cited Document
Table 5 represented the top 20 most published documents in the area of machine learning and deep learning application in healthcare. thus, the total citation and total citation per year citation of the top 5 documents in the application area of machine learning are Jiang F, 2017, Stroke Vasc Neurol (Total citation:467; TC per year:93.4), Rudin C, 2019, Nature of Mechanical Intelligence(Total citation:432; TC per year:144), Lundervold As, 2019, Journal of Medical Physics (Total citation:302; TC per year:101.6), Yu Kh, 2018, Nature of Biomedical Engineering (Total citation:268; TC per year:67), Chen M, 2017, IEEE Access(Total citation:255; TC per year:51). Similarly, in the area of deep learning application, the top most highly cited documents are Kermany DS, 2018, Cell (Total citation:862; TC per year:215.5), Coudray N, 2018, Nat Med (Total citation:542; TC per year:135.5), Rajkumar A, 2018, Npj Digit Med (Total citation:481; TC per year:120.25), Jiang F, 2017, Stroke Vasc Neurol (Total citation:470; TC per year:74), Esteva A, 2019, Nat Med (Total citation:461; TC per year:153.66). Apart from this, the total number of the cited document is highest Jiang F, 2017, Stroke Vasc Neurol (Total citation:467; TC per citation:101.6), while the maximum citation per year is Rudin C, 2019, Nat Mach Intell (total citation:432; TC per year:144) in machine learning in healthcare. Whereas, the maximum total citation and TC per year are Kermany DS, 2018, Cell(Total citation:862; TC per year:215.5) in the area of the deep learning application.
3.5 Top 20 Most Relevant Authors and their Production
The overall statistics of relevant authors state the numbers of authors significantly contributing to research areas. Thus, the total numbers of (5101/5916 authors) and (4871/6280 authors) authors are well contributing in the research field of application of machine learning and deep learning in the healthcare domain respectively. Likewise, out of the 5016 authors, single-authored documents are 16 and multi-authored documents are 5085, as well single-authored documents are 17 and multi-authored documents are 4854 in the application of deep learning and deep learning in the healthcare domain respectively. The analysis also reveals, larger groups of authors are contributing to the implementation of machine learning techniques in healthcare and producing more articles than deep learning application research area.
Table 6 demonstrated the top author's production and its fractional, thus, the top 5 authors Zhang Y(articles:14; article fractional:2.02), Li X(articles:11; article fractional:1.29), Li J(articles:10; article fractional:1.63), Li Y(articles:9; article fractional:1.49), Wang Z(articles:8; article fractional:1.39), similarly, Li H(articles:17; article fractional:2.12), Li J(articles:17; article fractional:2.34), Wang H(articles:14; article fractional:2.81), Yang J(articles:14; article fractional:2.15), Liu Y(articles:13; article fractional:2.21) in machine learning and deep learning in healthcare.
3.6 Top 20 Most Author Production Over Time
Figure 6 shows the year-wise top authors' production in the application of machine learning in healthcare. The diagram shows Zhang J is continuously working from the years 2011 -2021 and produced several articles more in the years 2020–2021. Authors Hussain A and Gupta S were involved in this research area from 2013 to, 2015 respectively. Rest authors Zhang Y, Li Y, Li X, Li J, Wang L, Chen Y, Hu YH, Chen M, etc. been involved in this area from the year 2015 and produced articles after the year 2020 on towards (Fig. 7)
Top author production over time in machine learning in healthcare
Figure 8 shows the year-wise top authors' production in the application of deep learning in healthcare. Deep learning techniques mostly came in research scenarios after the year 2015 and uses of its technique in healthcare mostly appeared after 2017.
Top author production over time in deep learning in healthcare
The diagram shows Zhang J is continuously working from the years 2011 -2021 and produced several articles more in the years 2020–2021. Authors Hussain A and Gupta S were involved in this research area from 2013 to, 2015 respectively. Rest authors Zhang Y, Li Y, Li X, Li J, Wang L, Chen Y, Hu YH, Chen M, etc. been involved in this area from the year 2015 and produced articles after the year 2020 on towards.
Li H, Wang H, Zhang Y, Li L, Wang Y, and Lee J has started contributing to research from the years 2017–2021. After the year 2018 Li J, Yang J, Liu Y, Muhammad G, Sun J, Hossain MS, Wang L, Xu Y, Zhang H, etc. started the application of deep learning in the area of the healthcare domain.
3.7 Top 20 Most Country Production
Table 7 presented the topmost country production, countries USA and China are heavily contributing to machine learning and deep learning application field research. In the USA the highest number of published articles is 2043 in the machine learning application field, while China is the maximum publication of 1103 articles in deep learning application research in the healthcare domain. Indian research heavily contributes to deep learning application(articles:330) implementation than machine learning(articles:231). The top 5 most country having maximum publications are the USA (Articles:2043), China (articles:589), the UK (articles:375), South Korea (articles:274), Canada(articles:264) and the USA (Articles:1327), China (articles:1103), South Korea(articles:389), India(articles:330), UK (articles:329) in machine learning and deep learning application in healthcare research respectively.
3.8 Top 20 Most Cited Country
The table demonstrates the top 20 most cited countries in the field of usages of machine learning and deep learning technique in the healthcare domain. And the citation of the country states the contribution to significant quality of research. hence, the top 5 country citation in the field of machine learning uses in the healthcare domain are the USA (citation:5072), China (citation:1434), United Kingdom (citation:847), India (citation:434), Norway (citation:316) and likewise in the field of deep learning uses in healthcare domain are USA (citation:5,625), China (citation:2,686), United Kingdom (citation:1,176), Korea (citation:904), Saudi Arabia (citation:601).
So, also tables indicate that countries the USA, China, and the UK contribute much to the application of machine learning and deep learning techniques in healthcare fields. But this three-country contribute the highest area of deep learning implementation in healthcare (Table 8).
3.9 Top 20 Most Affiliation Production
Here, Table 9 Shows the worldwide topmost 20 institutes, which have contributed more to research and published the highest article. Among all most 5 affiliations are Icahn School of Medicine at Mount Sinai, New York (articles:100), Harvard Medical School, Massachusetts (articles:78), Stanford University, California (articles:68), University of Toronto, Canada (articles:46), Pennsylvania State University (articles:34) and Stanford University, California (articles:109), King Saud University, Saudi Arabia (articles:89), Johns Hopkins University, Maryland (articles:84), Imperial College London, London (articles:47), Fudan University, Shanghai, China (articles:40) in uses of machine learning and deep learning technique in healthcare. Among all the top 20 affiliations, the highest articles produced by Stanford University, California in deep learning applications in healthcare also reveal the larger research in this research field.
3.10 Network Visualization
For visualization and network analysis of existing bibliometric information in the research area of machine learning, deep learning and healthcare VOS viewer software were used. This tool is applied for creating and representing a bibliometric science map of the author’s collaboration, country collaboration, institute collaboration, citation analysis, co-citation analysis, bibliographic coupling, co-word analysis, and co-authorship analysis as well as an article published by corresponding countries, etc. science map analysis pertains to the intellectual interaction and structural connection among research constituents [20]. Furthermore, text mining can be performed to construct and visualize co-occurrence networks of the most significant term extracted from the scientific literature by VOS viewer software.
Out of all bibliometric literature data downloaded from the WoS database in the research area of machine learning, deep learning, and healthcare, only 1000 literature records were selected for analyzing and visualization of the network from applying machine learning and deep learning in healthcare query keywords, that were extracted, and bibliometric analysis presented in this study. Article from this database was collected because this database contains only peer-reviewed articles [55].
3.11 Most Relevant Countries by Corresponding Author
Table 10, Fig. 9 and Fig. 10 represent the top 20 countries by corresponding author’s articles production. The top 5 countries USA, China, United Kingdom, India, and Korea have produced maximum and greatly contributed to the research area of the application of machine learning and deep learning technique in the advancement of the healthcare field. Thus, the total number of published articles is USA (345articles), China(101articles), United Kingdom (69 articles), India (60 articles), and Korea (52articles) and China (365 articles), USA (168 articles), Korea (68 articles), India (42 articles), United Kingdom (38 articles) in the application of machine learning and deep learning research fields.
Top 20 most Corresponding author’s country in the application of machine learning in the healthcare domain
Top 20 most Corresponding author’s country in the application of deep learning in the healthcare domain
Among all enlisted top 20 countries, the USA has significantly contributed with (263/345 articles) of SCP and (82/345 articles) of MCP, China with (70/101 articles) of SCP and (31/101 articles) of MCP, United Kingdom with (40/69 articles) of SCP and (29/69 articles) MCP, India with (50/60 articles) SCP and (10/60 articles) MCP, Korea with (38/52articles) SCP and (14/52articles) MCP. Similarly, China with (255/365 articles) in SCP and (110/365 articles) in MCP, USA has significantly contributed with (120/168 articles) in SCP and (48/168 articles) in MCP, Korea with (68/89articles) in SCP and (21/89 articles) MCP, India with (36/42 articles) SCP and (6/42 articles) MCP, United Kingdom with (21/38 articles) SCP and (17/38 articles) MCP. Overall, the USA, China, United Kingdom, Korea, and India have done huge research in this area and published maximum numbers of SCP, which shows that these countries have greatly contributed to the application area of machine learning in healthcare. While more numbers of MCP show worthy research collaboration with other countries.
3.12 Co-Authorship and Collaboration Analysis
Co-authorship collaboration analysis is a formal way of interaction among scholars. It is a most decent way to know, how scholars interact among themselves, including affiliation and countries associated with authors. Due to improving methodological and theoretical complexity in research, collaborations among scholars have become common, and collaborations among scholars contribute to showing greater clarity and richer insights between different scholars.
In Fig. 11, for building the co-authorship collaboration network between the authors the minimum number of documents of an author is 4 and the minimum number of citations of an author is 5, and out of 5528 authors, 17 meet the thresholds parameter were set. Then, a total of 17 authors were selected and associated in two clusters 1 and cluster 2. In this network, cluster1 represented in red colour has 3 group authors (de cecco, carlo n), (tesche, christian), (schoef, joseph) and green colour having 2 group authors (kim, young-hak), (lee, June-go). Among both cluster1 and cluster2, 4 group authors (tesche, christian), (schoepf, u. joseph), (de cecco, carlo n.) and (kim, young-hak) have equal numbers of link strength, and authors (lee, June-goo) have only one link connection in machine learning application areas.
Co-authorship visualization in machine learning application in healthcare
While, in Fig. 12, in deep learning application areas the most productive author was selected to meet threshold 8, out of 4220 authors, and it was associated with two clusters cluster1 and cluster2. In cluster1, the group authors are (Wang, Hao), (Zang, Yang), (Liu, Bo) and in cluster2 group, the authors are (Wang, Wei), (Huang, Wei). In this network, cluster1 is bigger than cluster2.
Co-authorship visualization in deep learning application in healthcare
3.13 Co-Authors-Country Collaboration Visualization
Figure 13 and Table 11 exhibited, the co-authors-country collaboration in the field of application of machines in the healthcare domain. For building the collaboration network, the parameter such as the minimum number of documents of a country 3 and the minimum number of citations of a country 5, and out of 85 countries, 51 meet the thresholds were taken. Based on the parameter 52 countries were selected and different colour size bubbles and link strength thickness of line, which shows the association among the co-authors-country interaction. Further, a total of 51 selected countries were associated in 7 clusters with 51 items.
Co-authors-Country collaboration visualization in machine learning application
Table 11 enlisted the top 20 countries having the highest number of published articles. Out of the 20 most countries, the top 5 countries, USA, England, peoples from China, India, and South Korea have a total of 410, 110, 97, 87, 69 articles, followed by a total citation of 6093, 1287, 1474, 686, 531 respectively. In this, the highest numbers of publications, citations and more numbers of strength shown by the USA, England, people from China, India, and South Korea have significantly contributed and a larger group of collaboration in this research field.
The Network (Fig. 14) has generated the minimum number of documents for country 5, and the minimum number of citations for country 10, to meet threshold 36. Out of the 77 countries, which meet the set-up parameter and most top 20 countries listed in Table 11 have the highest collaboration with each other. The maximum collaboration shown by countries are People from China (Article:358; Citation: 10,962), the USA (Article:249; Citation:11,291), South Korea (Article: 101; Citation: 2664), England (Article:66; Citation: 4015), India (Article:50; Citation: 579), Australia (Article:49; Citation: 2074), Canada (Article:43; Citation: 1937), have maximum numbers of collaboration with others countries and contributing most in deep learning application in the healthcare research area.
Co-authors-Country collaboration visualization in deep learning application
3.14 Co-Authors-Institute Collaboration Visualization
Table 12 and Fig. 15, show the top 20 most institute affiliation collaborations with the highest numbers of documents, citations, and total link strength.
Co-authors-Institute collaboration visualization in machine learning
For creating institute network collaboration, the minimum number of documents of an organization of 5, the minimum number of citations of an organization, and meet the threshold of 105 were set. Out of 1887 organizations, 105 institutes were selected. Therefore, the top 5 organizations, Harvard Medical School, Stanford University, Icahn School of Medicine at Mount Sinai, MIT, and King Saud University have the highest numbers of publications 42, 24, 19, 18, and 16 respectively. Although, the top 5 organizations Harvard Medical School, Stanford University, University of Oxford, University of Pennsylvania, and University Calif San Francisco have maximum citations 800, 573, 422, 371, 353, followed by link strengths 63, 27, 34, 23, 25 respectively. The more numbers of citations of organizations show the quality of research publication and the numbers of link strength show the numbers of collaboration. Therefore, selected organization publications and citations show that Harvard Medical School and Stanford University contribute to the better quality of research in the field of application of machine learning in the healthcare domain.
In Fig. 16, out of 1374 countries, the top 70 countries were selected to meet threshold 70 and countries having a minimum number of documents of an institute 5, minimum number of citations of an institute 10. Listed 20 countries, most published article and significant of research, top 5 countries Chinese Academy of Science, China (article:31; citation:896), Tsinghua University (article:18; citation:901), University Chinese Academy of Science (article:18; citation:539), Tsinghua University (article:18; citation:901), Stanford University (article:15; citation:546), National University of Defense Technology (article:14; citation: 1219), Harbin Institute of Technology (article:10; citation: 1607) in area of deep learning application in the healthcare domain.
Co-authors-Institute collaboration visualization in deep learning application
3.15 Occurrence of top 20 Most Author’s Keywords and Keyword plus
Table 13, Fig. 17, and Fig. 18 demonstrated the top 20 most occurrences of the author’s keywords and Keyword plus in the area application area of machine learning uses in the healthcare domain. For generating the network map, the parameter was used as the minimum number of occurrences of keyword 10 for all relevant keywords, and the threshold meets 38. Out of 2603 authors’ keywords, the most occurrence of 38 keywords was selected. Then, the top 10 most relevant author’s keywords in this field are “machine learning (Occurrence:650)”,” artificial intelligence (Occurrence:99)”,” healthcare (Occurrence:81)”,” deep learning (Occurrence:63)”,” big data (Occurrence:38)”,” classification (Occurrence:38)”,” covid-19(Occurrence:34)”,” internet of things (Occurrence:34)”,” natural language processing (Occurrence:34)”,” prediction (Occurrence:33)”. Among all 10 author’s keywords “machine learning (link Strength:614)”,” artificial intelligence (link Strength:157)”,” healthcare (link Strength:150)”,” deep learning (link Strength:115)”. Authors, enlisted keywords show that machine learning is the most popular word, which is applied in several fields. Likewise, for building a network map of the total numbers of authors’ keywords with higher numbers of occurrence frequency in deep learning applications in the healthcare domain, the threshold parameter 10, out of 2491 authors' keywords was set. Then, the top 10 highest generated authors' keywords are “deep learning (link Strength:762)”, “machine learning (link Strength:198)”, “artificial intelligence (link Strength:123)”, “neural network (link Strength:98)”, “training (link Strength:75)”, “convolutional neural network (link Strength:74)”, “classification (link Strength:67)”, “feature extraction (link Strength:62)”, “segmentation (link Strength:53)”, “Big Data (link Strength:45)”.
Network visualization of occurrence of top 20 most authors’ keywords in machine learning application
Network visualization of occurrence of the top 20 keywords plus in the machine learning area
As well, Table 13, Fig. 19, and Fig. 20 revealed the top 20 most occurrences of the most relevant Keywords plus which are popular in machine learning uses in the healthcare domain. For generating the network map, out of 2063 keywords plus, only 63 keywords were selected, onset parameter such as a minimum number of occurrences of keyword 10 for all relevant keywords plus, and the threshold meet 63. Further, among 63 selected keywords, the 10 most popular keywords in this field are “classification (Occurrence:113)”,” Prediction (Occurrence:82)”,” Risk (Occurrence:58)”,” diagnosis (Occurrence:57)”,” model (Occurrence:54)”,” validation (Occurrence:50)”,” Mortality (Occurrence:46)”,” Models (Occurrence:42)”,” System (Occurrence:42)”,” bigdata (Occurrence:40)”. Among all author’s keywords topmost popular keywords are “classification (link Strength:293)”,” Prediction (link Strength:181)”,” Risk (link Strength:112)”,” diagnosis (link Strength:122)”,” model (link Strength:126)”,” validation (link Strength:126)”,” Mortality (link Strength:105)”. Authors, enlisted keywords show the highest link strength with other relevant words, which is about other fields.
Network visualization of occurrence of top 20 most authors’ keywords in deep learning application
Network visualization of occurrence of top 20 most keywords plus keywords in deep learning application
Also, the selected total keywords plus, on the minimum number of occurrence frequency 10, meet thresholds 42, out of 1610 keywords were set up. Out of all generated 1610 keywords plus, the top 10 highest link strength keywords are” classification (link Strength:271)”,” Model (link Strength:177)”,” neural network (link Strength:176)”,” segmentation (link Strength:150)”,” algorithm (link Strength:123)”,” prediction (link Strength:102)”,” images (link Strength:95)”,” recognition (link Strength:92)”,” convolutional neural network (link Strength:81)”,” features (link Strength:73)”,” Diagnosis (link Strength:71)” (Figs. 21, 22, 23, 24)
Year-wise most trending authors’ keyword tree map of application of Machine learning in the healthcare domain
Year-wise most trending authors’ keyword tree map of application of Machine learning in the healthcare domain
Year-wise most trending keyword is Plus word cloud in Machine learning applied in the healthcare domain
Year-wise most trending keyword Plus word cloud in Deep learning applied in the healthcare domain
3.16 Year-Wise Identification of the top 20 most Trending Author’s KEYWORDS and Keywords in Machine Learning and Deep Learning uses in Healthcare.
The below table shows the top trending author's keywords and keywords plus keywords used in machine learning and deep learning application in the healthcare domain.
Table 14 represented the top 20 authors' keywords with years of frequency of machine learning techniques applied in the healthcare domain are “machine learning (freq.:617)”,”artificial intelligence (freq.:93)”, “healthcare (freq.:91)”, “deep learning (freq.:61)”,”prediction (freq.:43)” etc. likewise most trending authors keywords of deep learning techniques applying in healthcare domain are “deep learning (freq.:994)”,”machine learning (freq.:103)”, “artificial intelligence (freq.:65)”,”neural network (freq.:49)”,” Convolutional Neural Network (freq.:43)” etc. it also demonstrated the yearly most popular authors keyword are machine learning, artificial intelligence, healthcare, prediction model, data analytics, medical services, deep learning, diagnosis, Convolutional Neural Network, biomedical images, etc. in the application of machine learning and deep learning in healthcare fields during the year 2019–2021.
Table 15 exhibited yearly published numbers of the frequency with the popular keywords plus in the research area of machine learning and deep learning techniques applied healthcare domain. table enlisted the total numbers of an article with keywords plus classification(freq.:133), prediction (freq.:86), Risk (freq.:62), diagnosis (freq.:57), model (freq.:56), regression (freq.:26), segmentation (freq.:18) etc. and neural network(freq.:138), classification(freq.:123), model (freq.:78), algorithms (freq.:59), prediction(freq.:54), segmentation(freq.:50) etc. in area of Machine learning and deep learning in the healthcare domain.
These tables also demonstrate the year-wise highly popular keywords plus are classification, prediction, diagnosis, model, segmentation, classifier, data analytics, network, convolutional neural network, architecture, etc. in both the area of machine learning and deep learning applied in the healthcare field during the year 2018–2021. Among all listed keywords in Table15 classification, segmentation, classifier, network, model, and regression, the convolutional neural network is heavily used in the application filed research areas.
4 Summary
As emerging techniques, Machine Learning (ML) and Deep Learning (DL) have shown incredible potential in tackling challenging problems in several fields. This study has focused on those problems in healthcare that have been addressed using machine learning (ML) and deep learning (DL) with promising results globally. Both techniques have been shown as powerful tools in dealing with disease detection in preprocessing, feature extraction, feature selection, classification, and clustering steps. All literature was published on machine learning (ML) and deep learning (DL) in the Healthcare sector, as well as the application of machine learning (ML) and deep learning (DL) related to the healthcare domain.
This study has established the bibliometric analysis technique in the research area of machine learning, deep learning, and the healthcare field. And it revealed the worldwide research trends and performance analysis of the subject area. Up to now, there is a substantial gap in current research about the bibliometric analysis of Machine Learning (ML), Deep Learning (DL), and the Healthcare field. In this study, selected topical, title, and all fields’ keywords were used to extract the most relevant research paper from the Web of Science (WoS) core collection database, which included Science Citation Index Expanded (SCI) papers and articles, review article, early access document type paper from the period from 2010–30 June 2021.
Globally, in machine learning, deep learning, and healthcare, a total of 98,169, 51,559 and 1,67,326 articles with topical keywords search and 29,653, 19,493, and 38,109 articles with title-wise keywords search bibliometric information data were downloaded from the WoS database during 2010–2021(accessed 31 august 2021) respectively. Similarly, A total of 5,065, 2,755, and 4,588 articles from topic-wise categories and 1,380, 992, and 894 from title-wise categories with searched keywords Machine learning, Deep Learning, and Healthcare has been used in Indian scientific research. Additionally, exploring the application of machine learning in the healthcare and deep learning in the healthcare domain has published a total of 2,014 topics wise and 119 title-wise articles and 922 and 58 respectively worldwide. Similarly, an Indian research prospect in the application of machine learning and deep learning in the healthcare domain has published a total of 218 topic-wise articles and 16 title-wise articles globally and 142 articles topics-wise and 13 articles titles-wise at the Indian level (Table 1).
The purpose of this bibliometric data collection is to perform a bibliometric analysis, and network visualization, then evaluate the latest followed by the Document Type and Language, Publication output, Top Country Contribution, Top WoS core Categories and Journals, Top Authors, Top Research Areas and Analysis of Author Keywords, Keyword Plus related and most trending topic in machine learning, Deep learning uses in Healthcare. also, this paper analyses the number of citations based on the impact of multiple factors of the research paper as the number of authors, number of pages, and number of references. All bibliometric information data were downloaded by all relevant and latest literature published in ML, DL, and Healthcare research fields also implementation of machine learning and deep learning in healthcare fields, and all the source titles were included in web of science categories with three different document type article, review article, and early access article.
At the worldwide level, the total topical keywords-based article published is 86,292 research articles, 5,794 review articles, 2,854 early access articles with machine learning, 46,217 research article, 2,412 review article, 2,130 early access article in deep learning and 1,28,247 research articles, 22,071 review article, 3,540 early access article in the healthcare domain, also total numbers of research article 1,625, review articles 316, early access 98 in the application of machine learning in healthcare and total of 750 research articles, 149 review articles, 74 early access articles in deep learning application in healthcare. Comparatively, title wise query search was conducted and obtained a total of 22,935 research articles, 1,449 review articles and 924 with machine learning, followed by 16,055 research articles 730, review articles 788 early access articles in deep learning and 21, 545 research articles, 2,730 review article and 694 early access articles with healthcare, further, total research articles 74, review articles 11, early access articles 10 and 42 research articles, 5 review articles, 9 relay access articles in both machine learning and deep learning application in the healthcare sector (Table 2).
Also, worldwide five highest publication sources are IEEE Access, Journal of healthcare engineering, journal of biomedical informatics, PLOS One, BMC medical informatics, and decision-making publishing articles in machine learning and deep learning in healthcare fields. In the area of machine learning application these sources of publication total articles such as IEEE Access(articles:52), Journal of healthcare engineering(article:22), journal of biomedical informatics(article:20), PLOS One(article:20), BMC medical informatics, and decision making(article:19). Respectively, in deep learning in the healthcare area these sources of publication total articles like IEEE Access(articles:78), Journal of healthcare engineering(article:35), journal of biomedical informatics(article:15), BMC medical informatics and decision making(article:11).
From the global prospect, the top three countries that produced more research articles on the application of machine learning(ML) and deep learning(DL) in healthcare fields are the USA(articles frequency:2043), China(articles frequency:589), UK(articles frequency:375), South Korea (articles frequency:274), Canada(articles frequency:264), India (articles frequency:231) and USA(articles frequency:1327), China(articles frequency:1103), South Korea(articles frequency:398), India(articles frequency:330), UK (articles frequency:329). Among the top 10 most cited countries, the three topmost cited country having the USA (total citation:5072; avg articles citation:14.7), China (total citation:1434; avg articles citation:14.2), the UK (total citation:847; avg articles citation:12.28), India (total citation:434; avg articles citation:7.23), Norway (total citation:316; avg articles citation:63.2) in the area of machine learning use in healthcare. similarly, USA (total citation:5625; avg articles citation:26.41), China (total citation:2686; avg articles citation:13.57), UK (total citation:1176; avg articles citation:192.93), Korea (total citation:904; avg articles citation:12.05), Saudi Arabia (total citation:601; avg articles citation:13.07) in deep learning use in healthcare. Out of all the most cited and average cited countries USA has the highest total citations 1327 with an average article citation 26.41 in the use of deep learning in healthcare.
And top five authors are Zhang Y(articles:14), Li X(articles:11), Li J(articles:10), Li Y (article:9), Wang Z (article:8) and Li H(articles:17), Li J(articles:17), Wang H(aricles:14), Yang J(article:14), Liu Y(article:13) respectively.and the top five most affiliated Icahn School of Medicine at Mount Sinai, New York(articles:100), Harvard Medical School, Massachusetts(articles:78), Stanford University, California(articles:68) and Stanford University, California(articles:109), King Saud University, Saudi Arabia(articles:89), Johns Hopkins University(articles:84) are contributed greatly to both research fields.
The network analysis of the top 5 relevant terms yielded classification, prediction, neural network, model(s), design in machine learning, deep learning, quality, management, impact, services, and prevalence in healthcare research. During 2010–2021, the top 5 authors' keywords were machine learning, artificial intelligence, deep learning, classification, and neural networks in machine learning and deep learning, and healthcare, healthcare professionals, health policy, primary healthcare, patient safety, etc. in healthcare.
In addition, the top five trending keywords in the machine field are support vector machine, networks, algorithms, genetic algorithm approximation, neural network, classification, algorithm, architecture, dimensionality, sequence, mortality, attitudes, information, epidemiology, models, etc. The top authors' keywords in current research are classification, data mining, support vector machine(s), data analysis, unsupervised learning in machine learning, artificial intelligence, convolutional neural network, neural network, classification, object detection in deep learning, epidemiology, influenza, healthcare costs, healthcare service research, and e-health in healthcare.
5 Conclusion and Future Trends
When applied to any field of study, bibliometric analysis offers a fresh perspective on the study of research patterns. The purpose of this research was to apply bibliometric analysis techniques to uncover previously unknown relationships between machine learning, deep learning, and healthcare research areas, and to gain a deeper comprehension of how machine learning (ML) and deep learning (DL) methods are being used to improve healthcare delivery by minimizing human error in areas such as disease detection, diagnosis, prediction, drug discovery, precision medicine, robotic surgery, etc. This study provides a fresh perspective on the intersection of machine learning, deep learning, healthcare research, and the solutions they provide. In addition, the complete bibliometric information data on this topic has been analyzed using query keywords such as "machine learning," "deep learning," and "healthcare" from the web of science core collection from the period 2010 through August 30th, 2021, with a special emphasis on journals included in the SCI-Extended Index. In addition, this report discovers the extensive adoption of machine learning and deep learning techniques in healthcare and other domains by university researchers and industry practitioners. And it facilitates the application of these methods in other fields by researchers and practitioners.
From a global prospect, the top three countries that produced more research articles on the application of machine learning (ML), and deep learning (DL) in healthcare fields are the USA, China, and the UK. And top three authors are Zhang Y., Li X, Li J, Li H, Li J, Wang H and the top 3 most affiliated Icahn School of Medicine at Mount Sinai, New York(articles:100), Harvard Medical School, Massachusetts(articles:78), Stanford University, California (articles:68) and Stanford University, California(articles:109), King Saud University, Saudi Arabia(articles:89), Johns Hopkins University(articles:84) are contributed greatly in both research fields.
This study revealed that machine learning, and deep learning, in the healthcare field are some of the emerging areas of research that attract researchers to contribute to the future at a global level. Thus, the bibliometric study is a solution for detailed analysis of machine learning (ML) and deep learning (DL), and healthcare research areas and it helps to work in these areas special focusing on the healthcare domain. The key purpose of this endeavour is to raise awareness among academics and professionals about how machine learning (ML) and deep learning (DL) methods affect the medical field. More work needs to be done in this area, but it is expected to grow rapidly and have a greater effect on a structured dataset of greater size in the future. As a result, this ground-breaking study will assist scientists in maintaining a culture of constant innovation as they construct robust technologies tailored to the healthcare sector. We believe this research objective will encourage academics to keep exploring the potential of Machine learning and Deep learning in the medical field.
Data Availability
The datasets used and/or analyzed during the current study are available from the corresponding author upon reasonable request.
References
Garg M (2023) Mental health analysis in social media posts : a survey. Arch Comput Methods Eng. https://doi.org/10.1007/s11831-022-09863-z
Lundervold AS, Lundervold A (2019) An overview of deep learning in medical imaging focusing on MRI. Z Med Phys 29(2):102–127. https://doi.org/10.1016/j.zemedi.2018.11.002
Maity S, Bhattacharyya A, Singh PK, Kumar M, Sarkar R (2022) Last decade in vehicle detection and classification: a comprehensive survey. Arch Comput Methods Eng 29(7):5259–5296. https://doi.org/10.1007/s11831-022-09764-1
Magotra B, Malhotra D, Dogra AK (2022) Adaptive computational solutions to energy efficiency in cloud computing environment using VM consolidation. Arch Comput Methods Eng. https://doi.org/10.1007/s11831-022-09852-2
Gopal K, Swarnajit D, Rebika R, Arunita R (2023) Archimedes Optimizer : Theory, Analysis, Improvements, and Applications. Springer, Netherlands
Shamshirband S, Fathi M, Dehzangi A, Chronopoulos AT, Alinejad-Rokny H (2021) A review on deep learning approaches in healthcare systems: Taxonomies, challenges, and open issues. J Biomed Inform. https://doi.org/10.1016/j.jbi.2020.103627
Zhou S, Wu J, Zhang F, Sehdev P (2020) Depth Occlusion Perception Feature Analysis for Person Re-identification. Pattern Recognit Lett. https://doi.org/10.1016/j.patrec.2020.09.009
D. S. Maity, Niharika G, 2017 Machine Learning for Improved Diagnosis and Prognosis in Healthcare 1–9
Arooj A, Farooq MS, Akram A, Iqbal R, Sharma A, Dhiman G (2022) Big data processing and analysis in internet of vehicles: architecture, taxonomy, and open research challenges. Arch Comput Methods Eng 29(2):793–829. https://doi.org/10.1007/s11831-021-09590-x
Nemade V, Pathak S, Dubey AK (2022) A systematic literature review of breast cancer diagnosis using machine intelligence techniques. Arch Comput Methods Eng 29(6):4401–4430. https://doi.org/10.1007/s11831-022-09738-3
Caicedo D, Lara-Valencia L, Valencia Y (2022) machine learning techniques and population-based metaheuristics for damage detection and localization through frequency and modal-based structural health monitoring: a review. Arch Comput Methods Eng 29(6):3541–3565. https://doi.org/10.1007/s11831-021-09692-6
Wang X, Hu H, Liang Y, Zhou L (2022) On the mathematical models and applications of swarm intelligent optimization algorithms. Arch Comput Methods Eng 29(6):3815–3842. https://doi.org/10.1007/s11831-022-09717-8
Fathi M, Haghi Kashani M, Jameii SM, Mahdipour E (2022) Big data analytics in weather forecasting: a systematic review. Arch Comput Methods Eng 29(2):1247–1275. https://doi.org/10.1007/s11831-021-09616-4
Paturi UMR, Cheruku S, Reddy NS (2022) The role of artificial neural networks in prediction of mechanical and tribological properties of composites—a comprehensive review. Arch Comput Methods Eng 29(5):3109–3149. https://doi.org/10.1007/s11831-021-09691-7
Alhaqbani A, Kurdi HA, Hosny M (2022) Fish-inspired heuristics: a survey of the state-of-the-art methods. Arch Comput Methods Eng 29(6):3655–3675. https://doi.org/10.1007/s11831-022-09711-0
Chahar S, Roy PK (2022) COVID-19: a comprehensive review of learning models. Arch Comput Methods Eng 29(3):1915–1940. https://doi.org/10.1007/s11831-021-09641-3
Wamba MMQ, Fosso S (2021) Responsible artificial intelligence as a secret ingredient for digital health : bibliometric analysis, insights, and research directions. Inf Syst forntiers. https://doi.org/10.1007/s10796-021-10142-8
Kaur A, Singh Y, Neeru N, Kaur L, Singh A (2022) A survey on deep learning approaches to medical images and a systematic look up into real-time object detection. Arch Comput Methods Eng 29(4):2071–2111. https://doi.org/10.1007/s11831-021-09649-9
Aria M, Cuccurullo C (2017) bibliometrix: an R-tool for comprehensive science mapping analysis. J Informetr 11(4):959–975. https://doi.org/10.1016/j.joi.2017.08.007
N. J. Van Eck and L. Waltman, 2018 “VOSviewer Manual: Manual for VOSviewer version 1.6.7,” Univeristeit Leiden, 51
Jiang F et al (2017) Artificial intelligence in healthcare: past, present and future. STROKE Vasc Neurol 2(4):230–243. https://doi.org/10.1136/svn-2017-000101
Rudin C (2019) Models for high stakes decisions and use interpretable models instead. Nat Mach Intell 1(May):1–10
Yu K-H, Beam AL, Kohane IS (2018) Artificial intelligence in healthcare. Nat Biomed Eng 2(10):719–731. https://doi.org/10.1038/s41551-018-0305-z
Chen M, Hao Y, Hwang K, Wang L, Wang L (2017) Disease prediction by machine learning over big data from healthcare communities. IEEE ACCESS 5:8869–8879. https://doi.org/10.1109/ACCESS.2017.2694446
M. Matthew M Churpek, Trevor C Yuen, Christopher Winslow, David O Meltzer, Michael W Kattan, and Dana P Edelson, (2016) Multicenter comparison of machine learning methods and conventional regression for predicting clinical deterioration on the wards. Crit Care Med 44(2):368–374. https://doi.org/10.1097/CCM.0000000000001571
Nemati S, Holder A, Razmi F, Stanley MD, Clifford GD, Buchman TG (2018) An interpretable machine learning model for accurate prediction of sepsis in the ICU. Crit Care Med 46(4):547–553. https://doi.org/10.1097/CCM.0000000000002936
Wu J, Roy J, Stewart WF (2010) Prediction modeling using EHR data. Med Care 48(6):S106–S113. https://doi.org/10.1097/mlr.0b013e3181de9e17
Gianfrancesco MA, Tamang S, Yazdany J, Schmajuk G (2018) Potential biases in ML algorithms using EHR data. JAMA Intern Med 178(11):1544–1547. https://doi.org/10.1001/jamainternmed.2018.3763.Potential
Andrew Taylor MR, Pare JR, Venkatesh AK, Mowafi H, Melnick ER, Fleischman W, Kennedy Hall M (2016) Prediction of in-hospital mortality in emergency department patients with sepsis: a local big data–driven, machine learning approach. Acad Emerg Med 3(23):269–278. https://doi.org/10.1111/acem.12876
Narula S, Shameer K, Salem Omar AM, Dudley JT, Sengupta PP (2016) Machine-learning algorithms to automate morphological and functional assessments in 2D echocardiography. J Am Coll Cardiol 68(21):2287–2295. https://doi.org/10.1016/j.jacc.2016.08.062
Itu L et al (2016) A machine-learning approach for computation of fractional flow reserve from coronary computed tomography. J Appl Physiol 121(1):42–52. https://doi.org/10.1152/japplphysiol.00752.2015
Shameer K, Johnson KW, Glicksberg BS, Dudley JT, Sengupta PP (2018) Machine learning in cardiovascular medicine: are we there yet? Heart 104(14):1156–1164. https://doi.org/10.1136/heartjnl-2017-311198
Coenen A et al (2018) Diagnostic accuracy of a machine-learning approach to coronary computed tomographic angiography–based fractional flow reserve result from the MACHINE Consortium. Circ Cardiovasc Imaging 11(6):1–11. https://doi.org/10.1161/CIRCIMAGING.117.007217
Lalmuanawma S et al (2020) Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID- 19. The COVID-19 resource centre is hosted on elsevier connect, the company ’ s public news and information. Int J Adv Res Sci Commun Technol 4(January):160–164
Feng Z et al (2018) Machine learning-based quantitative texture analysis of CT images of small renal masses: differentiation of angiomyolipoma without visible fat from renal cell carcinoma. Eur Radiol 28(4):1625–1633. https://doi.org/10.1007/s00330-017-5118-z
Salvatore C, Cerasa A, Battista P, Gilardi MC, Quattrone A, Castiglioni I (2015) Magnetic resonance imaging biomarkers for the early diagnosis of Alzheimer’s disease: a machine learning approach. Front Neurosci. https://doi.org/10.3389/fnins.2015.00307
Wiens J, Shenoy ES (2018) Machine learning for healthcare: on the verge of a major shift in healthcare epidemiology. Clin Infect Dis 66(1):149–153. https://doi.org/10.1093/cid/cix731
Mohan S, Thirumalai C, Srivastava G (2019) Effective heart disease prediction using hybrid machine learning techniques. IEEE ACCESS 7:81542–81554. https://doi.org/10.1109/ACCESS.2019.2923707
Kermany DS et al (2018) Identifying medical diagnoses and treatable diseases by image-based deep learning. Cell 172(5):1122-1131.e9. https://doi.org/10.1016/j.cell.2018.02.010
Coudray N et al (2018) Classification and mutation prediction from non–small cell lung cancer histopathology images using deep learning. Nat Med 24(10):1559–1567. https://doi.org/10.1038/s41591-018-0177-5
Rajkomar A et al (2018) Scalable and accurate deep learning with electronic health records, npj Digit. Med 1(1):1–34. https://doi.org/10.1038/s41746-018-0029-1
Esteva A et al (2019) A guide to deep learning in healthcare. Nat Med 25(1):24–29. https://doi.org/10.1038/s41591-018-0316-z
Miotto R, Wang F, Wang S, Jiang X, Dudley JT (2017) Deep learning for healthcare: review, opportunities and challenges. Brief Bioinform 19(6):1236–1246. https://doi.org/10.1093/bib/bbx044
Lee J-G et al (2017) Deep learning in medical imaging: general overview. KOREAN J Radiol 18(4):570–584. https://doi.org/10.3348/kjr.2017.18.4.570
Faust O, Hagiwara Y, Hong TJ, Lih OS, Acharya UR (2018) Deep learning for healthcare applications based on physiological signals: a review. Comput Methods Programs Biomed 161:1–13. https://doi.org/10.1016/j.cmpb.2018.04.005
D. K. S. Nishijima David L, Wisner David H, Holmes James F (2016) Deep EHR: a survey of recent advances in deep learning techniques. Physiol Behav 176(5):139–148. https://doi.org/10.1109/JBHI.2017.2767063.Deep
Ting DSW et al (2019) Artificial intelligence and deep learning in ophthalmology. Br J Ophthalmol 103(2):167–175. https://doi.org/10.1136/bjophthalmol-2018-313173
Ignatov A (2018) Real-time human activity recognition from accelerometer data using convolutional neural networks. Appl Soft Comput J 62:915–922. https://doi.org/10.1016/j.asoc.2017.09.027
Schmidt-Erfurth U, Sadeghipour A, Gerendas BS, Waldstein SM, Bogunović H (2018) Artificial intelligence in retina. Prog Retin Eye Res 67(May):1–29. https://doi.org/10.1016/j.preteyeres.2018.07.004
Ravi D et al (2017) Deep learning for health informatics. IEEE J Biomed Heal Inform 21(1):4–21. https://doi.org/10.1109/JBHI.2016.2636665
Larson DB, Chen MC, Lungren MP, Halabi SS, Stence NV, Langlotz CP (2018) Performance of a deep-learning neural network model in assessing skeletal maturity on pediatric hand radiographs 1 PEDIATRIC IMAGING: neural network to assess skeletal maturity on pediatric hand radiographs larson et al materials and methods data acquisit. Radiology 287(1):313–322
Ortiz A, Munilla J, Górriz JM, Ramírez J (2016) Ensembles of deep learning architectures for the early diagnosis of the alzheimer’s disease. Int J Neural Syst 26(7):2–3. https://doi.org/10.1142/S0129065716500258
Steiner DF et al (2018) Impact of deep learning assistance on the histopathologic review of lymph nodes for metastatic breast cancer. Am J Surg Pathol 42(12):1636–1646. https://doi.org/10.1097/PAS.0000000000001151
Nagendran M et al (2020) Artificial intelligence versus clinicians: Systematic review of design, reporting standards, and claims of deep learning studies in medical imaging. BMJ 368:1–12. https://doi.org/10.1136/bmj.m689
Chakraborty S, Chakraborty S (2022) A scoping review on the applications of mcdm techniques for parametric optimization of machining processes. Arch Comput Methods Eng 29(6):4165–4186. https://doi.org/10.1007/s11831-022-09731-w
Author information
Authors and Affiliations
Contributions
Dr. Ela Kumar (EK) conceived and designed the study, Ms. Juli Kumari (JK) performed the research, analyzed the data, and Dr. Deepak Kumar (DK) contributed to editorial input. Conceptualization, methodology and formal analysis: JK, DK; investigation: EK; visualization: JK, DK; writing—original draft: JK; writing—review and editing: DK, EK; All authors read and approved the final manuscript.
Corresponding author
Ethics declarations
Conflicts of interest
The authors declare that there is no conflict of interests regarding the publication of this paper.
Ethical Approval
Not applicable.
Consent for Publication
All authors read and approved the final manuscript.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Kumari, J., Kumar, E. & Kumar, D. A Structured Analysis to study the Role of Machine Learning and Deep Learning in The Healthcare Sector with Big Data Analytics. Arch Computat Methods Eng 30, 3673–3701 (2023). https://doi.org/10.1007/s11831-023-09915-y
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11831-023-09915-y