1 Introduction

COVID-19 revealed several flaws and challenges that have been present in business for a lengthy duration, including how to perform everyday tasks from the standpoint of physical to virtual contacts. Companies may now determine which positions require face-to-face interaction and to what degree under the current operating model (Madero Gomez et al. 2020). This is an essential procedure because the workplace will undoubtedly grow from fundamental on-site to remote, face-to-face, or a mixture of these on the basis of how it is structured (Yoosefi Lebni et al. 2021a). For example, in hybrid or remote scenarios, like partially on location and partially remote, there arise advantages in the manner the workplace is intended for both the employees and the company, as it can permit them to work from their office, home or even a devoted remote office near to their homes, with lesser geographical conditions than entering the major headquarters office (Yoosefi Lebni et al. 2021b). As a result, people will live where they choose, which may be at a reduced cost of living or closer to loved ones, making the workplace a more appealing location to work. Attributing jobs to various workplace types can aid in determining who can work physically and who can work overseas (Risley 2020). This may be beneficial to both employers and workers, as it allows people to feel more at ease in their occupations and workspaces, as well as to be more inspired, leading to increased return for employers (Maqsood et al. 2021c) (NeJhaddadgar et al. 2020).

Human resource managers are seeking creative, innovative and efficient solutions to fix employee conflicts, sustain their health (Radic et al. 2020), and assistance them by constructing a dependable management scheme (Shuja et al. 2020). They are in responsible for hiring people, handling their effectiveness, benefits, and salaries as well as constructing and revamping employee categorizations. Considering the extraordinary and unpredictable conditions and shifting economic landscape, the historical issue of COVID-19 necessitates profound thought and adaptability in order to adequately handle human resources (Cooke et al. 2020). San Matthew Libraries' whole personnel was able to contact consumers via phone in even less than two months as a result of this crisis, and top teams were formed to find solutions and give services to the public in an entirely different and adaptable setting. They are also attempted to adapt to specific events inside the business by developing insightful inquiries and improving staff competencies (Su et al. 2021). Different environmental, political, and economic or health crises, such as COVID-19, should be utilized to examine and enhance the capacities of individuals and organizations (Fonseca and Azevedo 2020). Successful businesses all across the world, like Amazon, Facebook, and Google, have discovered that fast change provides incredible chances and benefits. COVID-19 motivates the researchers to undergo the deep learning process to detect the COVID- 19 patient rapidly. The existing method has classified the X-rays images into COVID- 19 and normal by using multi-model classification process (Verma et al. 2022). This multi-model classification incorporates Support Vector Machine (SVM) in the last layer of VGG16 Convolution network. For synchronization among VGG16 and SVM, they have added one more layer of convolution, pool, and dense between VGG16 and SVM. This study motivates us to find early stages tumors to increase the survival rate of the patient. This CAD system is efficient because it takes minimum time to detect whether the patient has cancer or not. This paper presents a system for finding the suspected reign for lung cancer using CT images. Image processing and ANN algorithms are used for this purpose. Mainly two ANN approaches are used. This method is further applied for different cancer detection (Nautiyal et al. 2019).

Every company works to meet its objectives through its personnel. Employers and workers must be able to have the ability to adapt to varying business demands in this challenging market. Employees face several challenges as a result of culture, technology, and varying work environments (Adam et al. 2021). To accomplish corporate goals, keep traditional workers, and recruit new workers, it's critical to put the entire employees' abilities and expertise to use. Due to the fact that employees are now working from home, firms are requiring greater levels of effectiveness from them. As employees must manage their job and family commitments, most times one among them will take priority over the other, resulting in an unbalance between home work and life in the current working environment (Albitar et al. 2020). Work-life balance is represented in the organization's effectiveness and performance. Scholars have agreed that in the face of massive health catastrophes, companies must be more alert and robust. The effect of the COVID-19 outbreak has elevated this problem to new heights, not just in China's economy but also globally (Arora and Suri 2020). COVID-19 has been spreading for more than a year, and its impact on HR challenges will persist for a long time. The major problem that organisations face today is not a singular incident, but a new reality that provides a variety of study opportunities for organisational researchers and practitioners. In the current situation, where often stated difficulties have been compounded by the emergence of COVID-19 and accompanied by an economic crisis, modern HR challenges can be rather diversified. Performance modification and staff well-being are examples of such issues (Bajrami and Demirovi et al. 2020); Employees cannot function as successfully as they could prior COVID-19, thus organisations are struggling to provide the same payment to them; trouble inspiring trust in employees (Bierema 2020); and heightened external and internal disputes amid an unforeseen vast-scale public crisis (Cai et al. 2020). What a broad spectrum of growing HR challenges, as well as the significant effects of the epidemic, would certainly necessitate more attention and debate.

After reviewing the literature works with many heuristic algorithms, it is observed that many limitations still need to be solved with an HRM on work from home scenario.

Q1. What are the aspects extracted from reviews?

Q2. What are the practical implications available for HRM on work from home scenario?

Q3. Which meta-heuristic algorithm can enhanced for improve the performance for the HRM on work from home scenario?

Q4. Why there is a need for optimizing the weight of DNN?

Q5. How to evaluate a HRM on work from home scenario using different performance metrics under evaluation?

The paper contribution is.

  • To conduct a thorough examination of HRD in IT firms in Tamil Nadu under a work-from-home scenario in the event of a COVID pandemic by gathering the responses of questionnaire from the HR in different IT companies of Tamil Nadu.

  • To develop a proposed DNN for the prediction process of responses, where the weight is optimized with the consideration of RMSE minimization.

  • To propose a novel form of optimization algorithm called ORS-DNN for enhancing the prediction process and to compare it with different methods for revealing its superiority in the HRM on work from home scenario due to the COVID pandemic situation.

  • To validate the efficient performance for the HRM on work from home scenario by comparing it with the conventional techniques by considering their performance measures.

The paper organization is: Section 1 provides the introduction of HRM on work from home scenario. The literature works are in Sect. 2. The proposed model of HRM on work from home scenario is given in Sect. 3. The employee wellbeing category is explained in Sect. 4. Section 5 describes the flexible workplace category. Section 6 elaborates on the remote work category. The job loss is discussed in Sect. 7. Section 8 describes the human capital category. The human resource development category is shown in Sect. 9. Section 10 elaborates the leadership category. Section 11 explains the performance category. The communication category is described in Sect. 12. Section 13 reveals the deep learning impact on HRM on work from home scenario during COVID pandemic situation. The results and discussions are provided in Sect. 14. The conclusion is in Sect. 15.

2 Literature survey

2.1 Related works

In 2021, Lucas et al. (2021) have looked at the existing state and future development of the workplace in the light of the COVID-19 pandemic. Documents, articles, and questionnaires from a variety of sources were reviewed to learn more regarding the individuals' and organisations' experiences with remote working, as well as the benefits and drawbacks of being able to get to work in the epidemic. The data analysis enabled the discovery of trends in the existing literature describing what had occurred and, in particular, its impact on the workplace. COVID-19's transformation has revolutionised the way organisations and workers work and will remain to do so, necessitating ongoing innovation of how they function and creating unprecedented actions, resulting in significant variations in the workplace. As a result, the workplace concept would never be what was envisaged prior to COVID-19, where work technology, reinvention and safety were major elements in the transformation.

In 2021, Reza et al. (2021) have recognised unforeseen problems, tactics, and uncommon decisions connected to human resource management in the COVID-19 pandemic other than clinical organizations. To mitigate COVID-19's effects, the research used a textual assessment technique primarily on the basis of businesses' human resource management practices. It used electronic databases such as PubMed,Web of Science, Scopus, LISTA, and PsycINFO to search for academic research. From November 2021 to the initial quarter of 2021, 1281 articles were extracted from the databases discussed. On the basis of the exclusion and inclusion criteria, this research evaluated chosen publications, identified 15 relevant articles, and deleted duplicates. Furthermore, on the basis of the results of the literature, the study built a conceptual framework for human resource management techniques to combat the COVID-19 pandemic.

In 2020, Thilagaraj (2020) have looked into a variety of issues that affected work-life balance when workers were required to work from home. The report's goal was to investigate the present state of work-life balance, and it discovered that work-life obligations had an influence on employees' personal lives. During the epidemic, the researchers gathered data from 60 individuals who worked from home. The data was investigated, and it was discovered that the employees' work-life balance was negatively impacted. As a result, the employees, who decide the company's success, should be provided with the ideal atmosphere in order to improve their job happiness.

In 2020, Gigauri (2020) has delivered a study that was dependent on qualitative expert conducting interviews. It highlighted the issues that HRM has faced in Georgia, as well as the repercussions of the catastrophe on human resources, the solutions that organizations have implemented, and advice for dealing with the situation from an HRM perspective. According to the conclusions of the expert interviews, organisations, in collaboration with HR managers, should build crisis management strategies and new rules for hybrid and remote working systems as a reaction to present and future problems.

In 2021, Zhong et al. (2021a) have undertaken a thorough literature study on developing research in the world of business and management to determine what the developing human resource difficulties were during the COVID-19 pandemic and offered relevant methods to address these concerns. Based on a review of the literature, nine major human resource concerns were identified across 13 sectors. COVID-19 has a huge influence on traditional human resource management, according to the result of this article, which necessitated researchers' empirical and theoretical attention. The proposals suggested relevant human resource methods for dealing with developing human resource difficulties, as well as a number of research locations for future studies in this topic.

In 2020, Caligiuri et al. (2020) have demonstrated the use of existing International Business (IB) research, with a focus on International HRM (IHRM), in addressing COVID-19 pandemic concerns. Multinational corporations' decision-makers have adopted a variety of measures to mitigate the pandemic's effects. Many of these efforts, whether at the macro or corporate level, have something to do with controlling distance and redefining limits. As the IB field was established as a valid area of academic investigation, managing distance as well as rethinking borders has been a fundamental concern of most of the IB research. Greater cross-border distance difficulties (e.g., as a consequence of travel prohibitions and limited international mobility) have resulted from the pandemic, as well as new intra-firm distancing concerns for formerly co-located staff. Distance poses challenges with respect to staff selection, support, training, health and safety, virtual collaboration, and leadership according to previous IHRM study. Most of this thought might be used to overcoming distance problems caused by pandemics. The current, exceptional examples of required physical distance really shouldn't appear to suggest equivalent rises in psychological distance, and also provided firms with just some knowledge into the unexpected advantages of a virtual workforce – a form of workforce that, indeed very possibly, will impact the post-COVID world's 'new normal.' Existing IHRM research did provide useful information for present, but there were still significant knowledge gaps. In terms of future IHRM research, three aspects were suggested: handling uncertainty, promoting global and international work, and rethinking organisational performance.

In 2021, Butterick and Charlwood (2021b) have revealed that the COVID-19 epidemic profounded labour market inequities. HRM ideas and practices that supported and legitimised the commercialization of labour were partly at fault for these inequities. During the epidemic, workers whose employment have been commercialised have harmed significantly. While HRM was not solely to blame for the misery, it was critical for those of us interested in studying, teaching, and practicing HRM to consider how our work has exacerbated a poor situation in order to improve in the future.

In 2020, Kahn et al. (2020) have begun to utilize telework (working from home) on a big scale due to the continuous economic and health problems caused by the COVID-19 pandemic and the mandatory physical distancing procedures. This might help to accelerate further use of teleworking methods even after the disaster, with a variety of consequences and unknown net impact on growth and various measures. S social partner cooperation and public policies were critical in ensuring that novel; welfare-improving and efficient working techniques created during the recession were preserved and established once the physical separation had ended. To enhance the welfare and productivity benefits of various widespread telework, governments should encourage investments in the managerial and physical capacity of workers and firms to telework, as well as label potential anxieties about worker well-being and long-term innovation, particularly associated with excessive workspace downscaling. How should public policy facilitate efficiency gains from teleworking in the post-COVID-19 era?

In 2020, Kaushik et al. (2020) have focused on the impact of COVID-19 virus pandemic on working life of employees. This research paper has also emphasized that how employers as well as HR managers were required to think out of the box and bring forth best practices as well as redefining HR roles during any adversity. It has also given light on few important issues such as People-Connect, adoption of a system of Skilling, re-skilling, Up-skilling and multi-skilling people about technology, design thinking, storytelling, analytics, Artificial intelligence to prepare our workforce to become more competent and talented by enhancing their skill set.

In 2021, Zito et al. (2021) have estimated a Structural Equations Model (SEM) for revealing that organizational communication was positively associated with self-efficacy and negatively with technostress and psycho-physical disorders. As mediators, technostress was positively associated with psycho-physical disorders, whereas self-efficacy was negatively associated. As regards mediated effects, results have showed negative associations between organizational communication and psycho-physical disorders through both technostress and self-efficacy. This study has contributed to literature underlying the role of communication in the current crisis and consequent reorganization of the working processes.

In 2021, Ayedee et al. (2021) have discussed various problems faced by human resource managers during the COVID-19 pandemic. The authors have seen a solution of the same in the application of emotional intelligence. The strategic human resource management approach can help HR managers in applying emotional intelligence. Strategic Human Resource Management (SHRM) was a proactive approach in comparison to traditional human resource management and it can help in formulating the policies effectively. The researchers have conducted online interviews to check the opinions of HR managers about the applicability of emotional intelligence.

2.2 Review

In all companies, the proliferation of COVID-19 causes disruption, uncertainty, complexity, and ambiguity. Any organization's most valuable asset is its people, who help it realise its objectives. As a result, the organizational strategy review is an acceptable response for managing human resources sustainably. Work-life balance is a constant challenge for all employees. The work-from-home scenario shifted the balance between job and personal life. Many prior studies have looked at the nature of work-life balance, but only a few have covered the effects of a pandemic on the workplace. Table 1 shows some of the features and challenges of existing impacts of work from home methods. Grounded theory (Antonio 2021) is flexible for employee care, family conciliation, and working and also paves the way for the feasibility of the variation of the organizations. But, it returns less timeline selection. Narrative review approach (Mohammad Reza Azizi 2021) is not used in any subjects and also enhances the employee health, welfare, adaptation to change, motivation, productivity, and satisfaction in the workplace. Still, further study is needed via questionnaires and interviews. WLB factor (Thilagaraj 2020) strengthens the communication of the organizations policies and also returns proper balance for spending the time with family. Yet, employees feel stressed regarding the overtime working. Qualitative expert interview research method (Gigauri 2020) does important achievement through the introduction of training employees and online HR processes and also achieves the outcomes through the adaptation to the new reality. But, the questionnaire is not modeled for the future quantitative survey. Emerging studies (Zhong et al. 2021a) solve the emerging HR issues for the enhancement of HRM research and also generate new problems by the novel HR practices. Still, the developed theories are not tested completely. Extant IB (Caligiuri et al. 2020) contains several features such as methodologically pluralist, multilevel, multi-stakeholder oriented, and multidisciplinary in nature and also supports and informs the management practice. Yet, it is impossible to answer the ‘big questions’ without the knowledge sharing. HRM (Butterick et al. 2021b) controls the ethical and moral beings with the help of a degree of agency and also reflects the paths for a better future. But, the case is not made for regulation and policy for enhancing the job quality. Telework (Kahn and Wiczer 2020) adapts the regulatory and legal system and also offers bilateral tax agreements. Still, it leads to hidden overtime. These challenges are lead to develop a novel method using the deep learning methods for controlling the issues generated in the work from home by the HRM.

Table 1 Features and challenges of existing impacts of work from home methods
Full size table

3 Proposed model of HRM on work from home scenario during COVID pandemic situation

3.1 Architectural representation

The goal of this study is to determine how COVID-19 affects the workplace and to investigate the presence of potential drivers that will allow for the analysis, study and monitoring of what the workplace of the future will be. How workplace transformation might occur in organisations in pandemic contexts like COVID-19 is of great relevance. At first, it appears that family reconciliation, remote work and flexible working schedules are deciding the workplace's longevity during the pandemic timespan; nevertheless, certain variables like passion at work, engagement and mindfulness may also be pertinent in this procedure, together with the COVID-19 (Soham 2021; Samajpaty 2021) context. The purpose of this study is to make a systematic analysis on HR in IT industries in Tamil Nadu during the work from home scenario in COVID pandemic situation. The analysis is conducted based on 9 different categories.

  • Employee Wellbeing

  • Flexible workplace

  • Remote work

  • Job loss

  • Human Capital

  • Human resource Development

  • Leadership

  • Performance

  • Communication

The architectural description of the introduced model of HRM on work from home scenario during COVID pandemic situation is shown in Fig. 1.

Fig. 1
figure 1

Developed Architecture of HRM on work from home scenario during COVID pandemic situation

Full size image

This study concentrates on three distinct phases.

  • In the first step, a questionnaire is created to address the issues that the above-mentioned drivers. The data collected for the questionnaire should be based on Tamil Nadu's IT HR viewpoint.

  • The second step entails gathering responses from various IT HR specialists in Tamil Nadu. As a result, the industrial authorities are being pressured to provide as much exact information as possible.

  • In the third step, prediction (Gupta et al. 2019) analysis is done using deep learning. Furthermore, combining deep learning with a meta-heuristic method aids in the long-term analysis.

4 Employee wellbeing of HR management on work from home scenario during Covid pandemic situation

This part describes the employee wellbeing of HR management on work from home scenario in the course of COVID pandemic situation. From Fig. 2a, the negative characteristics from the employees during the work from home is depicted as high demands by 9.5% of the HR, needs vast effort by 23.8% of HR, some consultation on variation by 33.3% of HR, role conflict by 23.8% of HR, and issues with remaining staff members by 9.5% of HR. On considering Fig. 2b, the employees with the positive characteristics in their job is shown as control over how you do or what you do it by 14.3% of HR, colleague support by 23.8% of HR, manager support by 42.9% of HR, and appropriate rewards by 19% of HR. In Fig. 2c, the employees handling the problems in positive manner during this work from home is described as concentrating on the problem and trying to handle it by 30% of HR, attaining social support by 35% of HR, and none by 30% of HR. while considering Fig. 2d, the problem handling with the employees in a passive manner during work from home in this pandemic situation is shown as avoiding them by 42.9% of HR, wishful thinking usage by 23.8% of HR, and blame yourself by 33.3% of HR. from Fig. 2e, the employees showing positive personality during work from home is shown as open by 14.3% of the HR, conscientious by 33.3% of HR, extravert by 14.3% of HR, agreeable by 33.3% of HR, and stable by 25% of HR. On considering Fig. 2f, the employees description during work from home in this pandemic situation is given as helping by 20% of HR, courteous by 30% of HR, and good sport by 50% of HR. In Fig. 2g, the employees present in the organization are comfortable during this work from home owing to the pandemic situation is depicted as high job satisfaction by 23.8% of HR, inspired employee who does not want to leave by 23.8% of HR, and none by 52.4% of HR. While considering Fig. 2h, the personal contract maintenance by the employees during this work from home situation is shown as maintaining promises by 28.6% of HR, treated fairly by 33.3% of HR, and high commitment by 38.1% of HR. From Fig. 2i, the employees having a high level of wellbeing satisfaction during this pandemic situation because of the work from home is shown as highly satisfied by 33.3% of HR, positive mood by 38.1% of HR, and happiness by 28.6% of HR.

Fig. 2
figure 2

Employee wellbeing of HR management on work from home scenario during COVID pandemic situation

Full size image

5 Flexible workplace of HR management on work from home scenario during Covid pandemic situation

This part shows the flexible workplace of HRM on work from home scenario during COVID pandemic situation. From Fig. 3a, maximum of the employees satisfied with the work from home option than the office during this pandemic situation is shown as strongly agree by 42.9% of HR, disagree by 28.6% of HR, neither agree nor disagree by 14.3% of HR, agree by 9.5% of HR and strongly disagree by 6% of HR. On considering Fig. 3b, employees requested for a variation in the work arrangements with respect to flexibility like workplace position, change in the core days or hours, and count of hours is given as yes by 28.6% of HR, no by 28.6% of HR, and prefers not to answer by 42.9% of HR. In Fig. 3c, the employees being loyal to the organization during this pandemic situation because of the work from home is shown as more by 14.3% of HR, less by 42.9% of HR, and no change by 42.9% of HR. While considering Fig. 3d, the employees looking for a job with another organization during this work from home owing to the inconvenient faced in this organization is given as yes by 28.6% of HR, no change by 38.1% of HR, and not sure by 33.3% of HR. From Fig. 3e, the employees intended on working for this company during this work from home is shown as yes by 28.6% of HR, no change by 38.1% of HR, and not sure by 33.3% of HR. On considering Fig. 3f, the company offering flexible working options to the employees during this work from home is shown as strongly agree by 23.8% of HR, disagree by 33.3% of HR, neither agree nor disagree by 23.8% of HR, agree by 14.3% of HR, and strongly disagree by 6% of HR. In Fig. 3g, the employees supporting in covering maximum turnover during this work from home condition is given as strongly agree by 14.3% of HR, disagree by 23.8% of HR, neither agree nor disagree by 33.3% of HR, agree by 23.8% of HR, and strongly disagree by 6% of HR.

Fig. 3
figure 3

Flexible workplace of HR management on work from home scenario during COVID pandemic situation

Full size image

6 Remote work of HR management on work from home scenario during Covid pandemic situation

This portion explains the remote work of HRM on work from home scenario during COVID pandemic situation. From Fig. 4a, the employees satisfied with the home-based workplace during this work from home is shown as yes by 23.8% of HR, no by 14.3% of HR, rarely by 14.3% of HR, often by 33.3% of HR, and prefer not to say by 14.3% of HR. On considering Fig. 4b, the employees working with more effectiveness during this work from home is shown as getting an opportunity to perform some other work by 19% of HR, not spending long duration in meetings by 28.6% of HR, and getting time to concentrate on their work by 52.4% of HR. In Fig. 4c, the controlling of employees on their working days during this work from home is given as yes by 9.5% of HR, no by 23.8% of HR, rarely by 23.8% of HR, often by 14.3% of HR, and prefer not to say by 28.6% of HR. While considering Fig. 4d, the flexibility for performing interviews with the remote people during this work from home is shown as yes by 14.3% of HR, no by 19% of HR, rarely by 28.6% f HR, often by 28.6% of HR, and prefer not to say by 9.5% of HR. From Fig. 4e, the missing of significant work tools with the employees during this work from home is shown as requires physical equipment by 19% of HR, requires documents and data by 38.9% of HR, and work tasks that cannot be done from home by 42.1% of HR. On considering Fig. 4f, the elaborate discussion of the meetings with the employees during this work from home is given as yes by 14.3% of HR, no by 9.5% of HR, rarely by 9.5% of HR, often by 38.1% of HR, and prefers not to say by 28.6% of HR. In Fig. 4g, remote work for the employees considered better than working from office is shown as strongly agree by 19% of HR, disagree by 19% of HR, neither agree nor disagree by 23.8% of HR, agree by 19% of HR, and strongly disagree by 19% of HR.

Fig. 4
figure 4

Remote work of HR management on work from home scenario during COVID pandemic situation

Full size image

7 Job loss of HR management on work from home scenario during Covid pandemic situation

This part explains the job loss of HRM on work from home scenario during COVID pandemic situation. From Fig. 5a, the reasons for not working or looking for work with the employees during this work from home is shown as pregnancy, injury, own illness by 19% of HR, personal family responsibilities by 33.3% of HR, training or education leave by 33.3% of HR, beginning to work later by 9.5% of HR, and no reason specified by 4% of HR. On considering Fig. 5b, the employees leaving the job due to the work pressure during this work from home because of the pandemic situation is given as yes by 9.5% of HR, No by 19% of HR, and prefer not to say by 71.4% of HR. In Fig. 5c, the employees ready for working in the company in the future amidst the work from home is shown as yes by 33.3% of HR, no by 28.6% of HR, and prefer not to say by 38.1% of HR. While considering Fig. 5d, the employees preferring training for avoiding the loss of job in the future due to this work from home is shown as yes by 19% of HR, no by 23.8% of HR, and prefer not to say by 57.1% of HR. From Fig. 5e, the employees working in a usual manner to avoid the loss of job during this work from home is given as others else with pay by 19% of HR, others without pay by 6% of HR, family gain by 23.8% of HR, cooperative member by 42.9% of HR, and other by 9.5% of HR. On considering Fig. 5f, the major obstacle in getting a job for the employees during this work from home is shown as no education by 14.3% of HR, inappropriate general education by 28.6% of HR, inappropriate vocational education by 28.6% of HR, in accurate training opportunities by 23.8% of HR, and improper work experience by 6% of HR. In Fig. 5g, the employees resent and actively viewing for work before getting this current job during this pandemic situation is given as less than a week by 9.5% of HR, 1–4 weeks by 28.6% of HR, 1–2 months by 33.3% of HR, 3–6 months by 19% of HR, and 6 months–1 year by 9.5% of HR.

Fig. 5
figure 5

Job Loss of HR management on work from home scenario during COVID pandemic situation

Full size image

8 Human capital of HR management on work from home scenario during Covid pandemic situation

This section explains the human capital of HR management on work from home scenario during COVID pandemic situation. From Fig. 6a, the data presently gathered from the employees for assessing the leadership and management capability during this work from home is shown as employee demographics by 14.3% of HR, management ability by 14.3% of HR, development and training quality by 47.6% of HR, absence by 19% of HR, and recruitment by 5% of HR. On considering Fig. 6b, the link among organizational performance and employee metrics during this work from home is given as yes by 14.3% of HR, no by 38.1% of HR, and never by 47.6% of HR. In Fig. 6c, the key employee indicator incorporation into the monitoring of entire performance of the organization during this work from home is shown as balanced scorecard by 23.8% of HR, management board reports by 28.6% of HR, and others by 47.6% of HR. While considering Fig. 6d, single or very small key employee indicator count in this pandemic situation is given as yes by 19% of HR, no by 28.6% of HR, and often by 52.4% of HR. From Fig. 6e, the data containing highest significance for understanding the employee-performance link in this pandemic situation due to the work from home is given as recruitment by 14.3% of HR, retention by 28.6% of HR, talent management by 42.9% of HR, and reward by 14.3% of HR. On considering Fig. 6f, the employee management metrics being gathered in a routine manner in this pandemic situation is shown as leadership ability by 19% of HR, development and training investment by 19% of HR, engagement/staff motivation by 33.3% of HR, retention by 19% of HR, and performance by 9.5% of HR. In Fig. 6g, the different metrics for gathering the data from the employees during this work from home is given as human investment ratio by 19% of HR, total employment costs by 9.5% of HR, resignations by 33.3% of HR, commitment index by 23.8% of HR, and remuneration by 14.3% of HR.

Fig. 6
figure 6

Human Capital of HR management on work from home scenario during COVID pandemic situation

Full size image

9 Human resource development of HR management on work from home scenario during Covid pandemic situation

This part explains the human resource development of HRM on work from home scenario during COVID pandemic situation. From Fig. 7a, the HR having complete aware of the business strategies and requirements in this pandemic situation is shown as fully agree by 23.8% of HR, agree by 14.3% of HR, neither agree nor disagree by 23.8% of HR, partially disagree by 23.8% of HR, and fully disagree by 14.3% of HR. On considering Fig. 7b, the efforts considered for creating awareness within the employees regarding the customer requirements, financial position, cost, quality of service/product of the organization during this work from home is shown as fully agree by 28.6% of HR, agree by 14.3% of HR, neither agree nor disagree by 28.6% of HR, partially disagree by 14.3% of HR, and fully disagree by 14.3% of HR. In Fig. 7c, the emphasis of the organization on each of the below recruitment sources in this COVID pandemic is shown as printing advertisements by 28.6% of HR, career/Internet sites by 14.3% of HR, consultants/placement agencies by 28.6% of HR, educational institutes by 14.3% of HR, and employee referrals by 14.3% of HR. While considering Fig. 7d, the employee contract type preferred by the organization during this work from home is given as permanent by 9.5% of HR, contractual by 33.3% of HR, temporary by 33.3% of HR, and part time by 23.8% of HR. From Fig. 7e, the significance to several issues that are associated with organizational workflows during this work from home with the employees is given as efficiency by 19% of HR, creativity and innovation by 19% of HR, flexibility by 23.8% of HR, wide job classes by 23.8% of HR, and detailed work planning by 14.3% of HR. On considering Fig. 7f, the degree of usage with consideration of selecting the employees in this COVID pandemic is shown as psychological tests by 14.3% of HR, performance tests by 19% of HR, realistic job previews by 23.8% of HR, trainability by 28.6% of HR, and team-oriented selection by 14.3% of HR. In Fig. 7g, the organization having a broad network of computerized human resource information system along with the latest software in this work from home is given as fully agree by 23.8% of HR, agree by 19% of HR, neither agree nor disagree by 28.6% of HR, partially disagree by 23.8% of HR, and fully disagree by 4% of HR.

Fig. 7
figure 7

Human Resource Development of HR management on work from home scenario during COVID pandemic situation

Full size image

10 Leadership of HR management on work from home scenario during Covid pandemic situation

This part describes the leadership of HRM on work from home scenario during COVID pandemic situation. From Fig. 8a, the maintenance of leadership perfectly during this work from home is given as not true by 19% of HR, seldom true by 28.6% of HR, occasionally true by 19% of HR, somewhat true by 23.8% of HR, and very true by 9.5% of HR. On considering Fig. 8b, allocating and obtaining resources seems to be a challenging aspect of leadership job during this pandemic situation is given as not true by 9.5% of HR, seldom true by 23.8% of HR, occasionally true by 33.3% of HR, somewhat true by 23.8% of HR, and very true by 9.5% of HR. In Fig. 8c, the successful conflict resolution is the leadership opponent respect in this work from home scenario is shown as not true by 6% of HR, seldom true by 28.6% of HR, occasionally true by 38.1% of HR, somewhat true by 14.3% of HR, and very true by 14.3% of HR. While considering Fig. 8d, flexible leadership that makes variations in the organization in this pandemic situation is given as fully agree by 23.8% of HR, agree by 14.6% of HR, neither agree nor disagree by 28.6% of HR, partially disagree by 19% of HR, and fully disagree by 14.3% of HR. From Fig. 8e, the employees to be considered as a portion of the decision making process in this work from home is given as strongly disagree by 23.8% of HR, disagree by 14.3% of HR, neutral by 28.6% of HR, agree by 19% of HR, and strongly agree by 14.3% of HR. On considering Fig. 8f, the leadership needs staying apart from the employees as they perform their work in this pandemic situation is given as fully agree by 14.3% of HR, agree by 19% of HR, neither agree nor disagree by 38.1% of HR, partially disagree by 14.3% of HR, and fully disagree by 14.3% of HR. In Fig. 8g, the leaders offering employees total freedom for handling problems on their self in this work from home condition is shown as strongly disagree by 14.3% of HR, disagree by 19% of HR, neutral by 28.6% of HR, agree by 23.8% of HR, and strongly agree by 14.3% of HR.

Fig. 8
figure 8

Leadership of HR management on work from home scenario during COVID pandemic situation

Full size image

11 Performance of HR management on work from home scenario during Covid pandemic situation

This section describes the performance of HRM on work from home scenario during COVID pandemic situation. From Fig. 9a, the special achievement or work success from the employees during this work from home is given as yes by 42.9% of HR, no by 19% of HR, and prefer not to say by 38.1% of HR. On considering Fig. 9b, the employee performance greater on the job during this pandemic situation is shown as all of the time by 19% of HR, most of the time by 23.8% of HR, some of the time by 33.3% of HR, a little of the time by 14.3% of HR, and none of the time by 9.5% of HR. In Fig. 9c, the employees not focusing on their job during this COVID situation is given as all of the time by 14.3% of HR, most of the time by 28.6% of HR, some of the time by 19% of HR, a little of the time by 33.3% of HR, and none of the time by 5% of HR. While considering Fig. 9d, the health problems lessening the amount or kind of work the employees perform because of the COVID breakthrough is shown as all of the time by 14.3% of HR, most of the time by 19% of HR, some of the time by 33.3% of HR, a little of the time by 19% of HR, and none of the time by 14.3% of HR. From Fig. 9e, the normal job performance rating of the employees over these two years because of the work from home due to the COVID pandemic is shown as zero by 19% of HR, two by 19% of HR, five by 19% of HR, seven by 23.8% of HR, and ten by 19% of HR. On considering Fig. 9f, the comparison of the entire job performance of one employee with the performance of various employers having a similar form of job in this COVID pandemic is shown as much better by 9.5% of HR, somewhat better by 19% of HR, average by 19% of HR, somewhat worse by 33.3% of HR, and lot worse by 19% of HR. In Fig. 9g, the employees considering challenging tasks for improving their performance during this work from home situation is given as yes by 19% of HR, no by 38.1% of HR, and often by 42.9% of HR.

Fig. 9
figure 9

Performance of HR management on work from home scenario during COVID pandemic situation

Full size image

12 Communication of HR management on work from home scenario during Covid pandemic situation

This part explains the communication of HRM on work from home scenario during COVID pandemic situation. From Fig. 10a, communication with the employees goes better during this COVID pandemic is shown as never by 28.6% of HR, not often by 28.6% of HR, sometimes by 19% of HR, often by 14.3% of HR, and always by 9.5% of HR. On considering Fig. 10b, effective communication attained easier by considering turns talking with the employees during this work from home is given as never by 14.3% of HR, not often by 33.3% of HR, sometimes by 23.8% of HR, often by 19% of HR, and always by 9.5% of HR. In Fig. 10c, the employees making perfect communication during the meetings in this work from home scenario is shown as yes by 9.5% of HR, no by 14.3% of HR, rarely by 14.3% of HR, very often by 52.4% of HR, and often by 9.5% of HR. While considering Fig. 10d, the communication of employees with the organization during this work from home is given as very good by 9.5% of HR, good by 9.5% of HR, fair by 28.6% of HR, poor by 28.6% of HR, and very poor by 23.8% of HR. From Fig. 10e, the employees suffering from the listening problems in this work from home situation is shown as unfamiliar with the foreigner’s pronunciation and accents by 19% of HR, not able to catch the conversations or words by 23.8% of HR, not able to get the technical terms by 19% of HR, not able to remember the complete information by 23.8% of HR, and others by 14.3% of HR. On considering Fig. 10f, the company improving the English communication skills of the employees during this work from home situation is shown as using English during work by 9.5% of HR, considering English courses at a language institute by 42.9% of HR, training abroad by 19% of HR, in-house training by 23.8% of HR, and others by 5% of HR. In Fig. 10g, the communication affecting the employees in the organization during this pandemic situation because of the work from home is given as to some extent by 14.3% of HR, to vast extent by 38.1% of HR, never by 23.8% of HR, and prefers not to say by 23.8% of HR.

Fig. 10
figure 10

Communication of HR management on work from home scenario during COVID pandemic situation

Full size image

13 Deep learning impact on HRM on work from home scenario during COVID pandemic situation

13.1 Deep learning for HRM on work from home scenario

Here, nine categories are described for analyzing the HRM on work from home scenario during COVID pandemic situation. Every category is composed of more than seven questions. The DNN network is assessed for every category, thus a total of 9 DNN networks are assessed. Every question contains responses that are designed in the format of multiple responses. The responses are gathered from the HR IT professionals of Tamil Nadu. The response changes on the basis of every preference of the HR. From the collected responses, assigning score for every response takes place. The possibilities associated with the score of each response are defined. These possibilities are subjected to the proposed DNN for the future response prediction. The training and the testing process occur in the proposed DNN. In the training process, the possibilities are subjected as input to the proposed DNN, and the target is considered fixed that is the mean (scores). During the testing phase, the same possibilities are provided as input to the proposed DNN and the output obtained is the predicted (score), which is associated to the future responses. Here, the weight of the DNN is optimized by the suggested ORS-TSA, thus known as proposed DNN. The overall response is predicted by this proposed DNN, which is described as the predicted response that is called as the output. The architectural model of the deep learning effect on HRM on work from home during COVID pandemic situation is displayed in Fig. 11. This, in turn provides a clear depiction on distinct HRM for work from home scenario.

Fig. 11
figure 11

Architectural model of the deep learning effect on HRM for work from home scenario

Full size image

13.2 Proposed DNN

The DNN is considered as the learning algorithm for analyzing the HRM on work from home scenario during COVID pandemic situation. The three major components of DNN (Ramesh et al. 2019a) are the "input layer," "output layer," and "hidden layers." It contains two hidden layers that aid in the mapping of "output and input data." Due to the rising number of training iterations, it is still difficult to satisfy the decision limit. "Training speed and classification accuracy" is improved by constructing two hidden layers. Cross entropy is used as the loss function of DNN as the preparation for training and testing. The use of cross-entropy losses greatly improved the performance of the sigmoid and soft max output models. The optimization of DNN is to improve the performance of the network. It can be used to achieve the best performance by reducing the errors. The optimization of significant parameters in each technique is performed through the optimization algorithm for getting the optimal solutions. Here, the trial and error method is used for assigning the range for particular constraints, where the range is decided based on the optimal solutions attained in the training phase. The total node count is determined as in Eq. (1) when the hidden layer is taken into consideration.

$$nw = \sqrt {aw + bw} + cw$$
(1)

The input layer node count is supplied by \(aw\), the output layer node count is supplied by \(bw\), the hidden layer node count is supplied by \(nw\), and a constant value is supplied by \(cw\). In the case of hidden layer, non-linear fitness capacity is used as an activation function. The sigmoid is utilized as an activation function, as demonstrated in Eq. (2).

$$SW = \frac{1}{{1 + e^{ - resw} }}$$
(2)

Here, Eq. (3) provides the input data \(resw\), and the mapping function \(MW_{fw}\) is used to activate it.

$$MW_{fw} = sigm\left( {\omega_{iw} resw + \beta_{iw} } \right)$$
(3)

The bias is denoted by \(\beta\), whereas the weight matrix is denoted by \(\omega\).

The DNN provides advantages such as well for various applications, sophisticated capabilities, ultimate platform adaptability, and so on. However, it is URL-dependent, requires a large amount of data, and increases the cost to consumers, and so on. To address these flaws, the weight of DNN is optimized, resulting in proposed DNN. This proposed DNN has advantages such as improved security, ease of usage, and dependability. The weight of DNN has a boundary range of 5–255. As a result, the objective is shown in Eq. (4).

$$Obj = \mathop {\arg \min }\limits_{{\left\{ {WT_{DNN} } \right\}}} \left( {RMSE} \right)$$
(4)

The objective is defined by \(Obj\) and the weight of proposed DNN is specified by \(WT_{DNN}\) accordingly. "RMSE is a commonly used measure of the discrepancies between values predicted by a model or an estimate and the values observed," according to Eq. (5).

$$RMSE = \sqrt {\frac{{\sum\limits_{eg = 1}^{ng} {\left( {fg_{eg2} - ag_{eg1} } \right)^{2} } }}{ng}}$$
(5)

The computed value for each fitted point is represented by \(eg\), the actual value by \(ag\), the fitted point count by \(ng\), and the forecast value by \(fg\) in the above equation, respectively.

13.3 Proposed ORS-TSA

The proposed ORS-TSA for the HRM on work from home scenario during the COVID pandemic situation enhances the prediction phase through the optimization of the weight of the DNN with the consideration of RMSE minimization. TSA (SatnamKaur et al. 2020) mimics the behaviours of "jet propulsion and tunicate swarm." It's a solution approach for "non-linear constrained issues." Tunicate swarm activity triggers it for living in the deep ocean. To statistically simulate jet propulsion behaviour, a tunicate must meet three criteria: avoid conflicts among search agents, go along the path of the best search agent's position, and stay close to the optimum search agent. Swarm activity will keep a number of different search agents informed on the optimum response. To mathematically model the jet propulsion behaviour, a tunicate should satisfied three conditions namely avoid the conflicts between search agents, movement towards the position of best search agent, and remains close to the best search agent. The efficiency of this algorithm is further compared with several well-regarded met heuristic approaches based on the generated optimal solutions. The simulation results demonstrate that TSA generates better optimal solutions in comparison to other competitive algorithms and is capable of solving real case studies having unknown search spaces.

Conflicts within the search agents are avoided by computing fresh search agent placement with the help of a vector \(\vec{E}\), as detailed below.

$$\vec{E} = \frac{{\vec{K}}}{{\vec{Q}}}$$
(6)
$$\vec{K} = g_{2} + g_{3} - \vec{J}$$
(7)
$$\vec{J} = 2 \cdot g_{1}$$
(8)

The social pressures inside the search agents are represented by \(\vec{Q}\), random numbers by \(g_{1}\), \(g_{2}\), and \(g_{3}\), water flow advection is represented by \(\vec{K}\), and gravitational force is represented by \(\vec{J}\) in the aforementioned equations. The vector \(\vec{J}\) is computed using Eq. (9).

$$\vec{Q} = \left\lfloor {T_{\min } + g_{1} \cdot T_{\max } - T_{\min } } \right\rfloor$$
(9)

The starting speed is equal to \(T_{\min }\) and the subordinate speed is equal to \(T_{\max }\) in this situation. As in Eq. (10), the search agents are then relocated in the direction of the best neighbour.

$$P\vec{H} = \left| {F\vec{W} - rd \cdot \vec{T}_{t} \left( b \right)} \right|$$
(10)

The random number is represented by \(rd\), the tunicate position is represented by \(\vec{T}_{t} \left( b \right)\), the food source position is represented by \(F\vec{W}\), the current iteration is represented by \(b\), and the distance between the search agent and the food source is represented by \(P\vec{H}\).

As in Eq. (11), the search agent maintains his place in the route of the best search agent.

$$\vec{T}_{t} \left( {b^{\prime}} \right) = \left\{ {\begin{array}{*{20}c} {F\vec{W} + \vec{E} \cdot P\vec{H}} & {if\;rd \ge 0.5} \\ {F\vec{W} - \vec{E} \cdot P\vec{H}} & {if\;rd < 0.5} \\ \end{array} } \right.$$
(11)

Clearly, the term \(\vec{T}_{t} \left( {b^{\prime}} \right)\) describes the updated tunicate position with reference to the food supply \(F\vec{W}\). The top two optimal best solutions are kept, and other search agent locations are adjusted on the basis of the best search representative position, to statistically mimic tunicate's swarm behaviour. The tunicate swarm behaviour is characterized by the formula below, which is similar to Eq. (12).

$$T_{t} \left( {\vec{b} + 1} \right) = \frac{{\vec{T}_{t} \left( b \right) + T_{t} \left( {\vec{b} + 1} \right)}}{{2 + g_{1} }}$$
(12)

The TSA has various advantages, like improved convergence, global optimum solutions, reduced vulnerability, increased scalability and sensitivity, and the capability to apply it to real-world case scenarios. Unfortunately, it is incapable of solving "binary or multi-objective optimization problems." As a result, worst position is also considered through the random search to improve performance, resulting in the algorithm known as ORS-TSA. This ORS-TSA reduces computing time and simplifies the solution of "binary and multi-objective optimization problems."

In the traditional TSA, the random number \({\text{g}}_{{2}}\) lies in the range of [0,1]. But, in the proposed ORS-TSA, the term \({\text{g}}_{{2}}\) is computed with the help of oppositional random search-based concept as below in Eq. (13).

$${\text{g}}_{{2}} = \frac{{Obj_{worst} - Obj_{best} }}{{Obj_{mean} }}*\vec{Q}$$
(13)

Here, the worst fitness is given by \(Obj_{worst}\), best fitness is given by \(Obj_{best}\), and the mean fitness is given by \(Obj_{mean}\) respectively. The pseudo code of ORS-TSA is shown in Algorithm 1 and its flowchart is given in Fig. 12.

figure a
Fig. 12
figure 12

Flowchart of proposed ORS-TSA

Full size image

14 Results and discussions

14.1 Experimental setup

The proposed model of HRM on work from home scenario during COVID pandemic situation was implemented in MATLAB 2020a and the outcomes were analyzed. The population size and the iteration count were 10 and 25. Here, the proposed ORS-TSA-DNN was compared with various heuristic-based algorithms like PSO-DNN (Humam et al. 2019b), GWO-DNN (Bansal and Singh 2021), WOA-DNN (Dinakara et al. 2017), and TSA-DNN (SatnamKaur et al. 2020) as well as classifier algorithms like NN (Huang et al. 2001), KNN (Anushya et al. 2019), SVM (Wang et al. 2009), and DNN (Ramesh et al. 2019a) in terms of several error measures to describe the superiority of the proposed method. The hidden neuron count for the proposed approach can be listed below, for category 1 is [70,255], category 2 is [100,212], category 3 is [40,167], category 4 is [133,216], category 5 is [70, 30], category 6 is [141,254], category 7 is [158,160], category 8 is [227, 99], category 9 is [139, 80]. The hidden neuron count for the neural network is 10. Finally, the hidden neuron count for DNN is (Adam et al. 2021; Adam et al. 2021). The whole dataset has been divided into two parts, where 70% of data have been used for training and 30% of data have been used for testing. Thus, all the data have been trained and tested efficiently to show the performance of the proposed model.

14.2 Error metrics

The different error metrics are described below (Chicco et al. 2021d).

14.3 RMSE analysis in heuristic basis

The RMSE analysis of different categories of the proposed HRM on work from home during COVID pandemic situation is given in Fig. 13. From Fig. 13d, for job loss, the RMSE of ORS-TSA-DNN at 65% learning percentage is 10.34%, 7.14%, 3.70%, and 13.33% higher than TSA-DNN, WOA-DNN, GWO-DNN, and PSO-DNN respectively. On considering Fig. 13e, for human capital, the RMSE of ORS-TSA-DNN at 35% learning percentage is 42.86, 33.33, 20, and 50% surpassed than TSA-DNN, WOA-DNN, GWO-DNN, and PSO-DNN respectively. In Fig. 13f, for human resource development, the RMSE of ORS-TSA-DNN at 45% learning percentage is 17.65, 12.5, 6.67, and 22.22% improved than TSA-DNN, WOA-DNN, GWO-DNN, and PSO-DNN respectively. Thus, the RMSE analysis is better with the ORS-TSA-DNN than the other heuristic-based algorithms for the HRM on work from home scenarios due to the COVID pandemic situation.

Fig. 13
figure 13

RMSE analysis of different heuristic-based algorithms for the HRM on work from home scenario with respect to, a Employee Wellbeing, b Flexible workplace, c Remote Work, d Job loss, e Human capital, f Human resource development, g Leadership, h Performance, and i Communication

Full size image

14.4 RMSE analysis in classifier basis

The RMSE analysis of different classifier algorithms for the HRM on work from home scenario is shown in Fig. 14. From Fig. 14a, for employee wellbeing, the RMSE of ORS-TSA-DNN at 45% learning percentage is 33.33, 30, 26.32, and 22.22% more than DNN, SVM, KNN, and NN respectively. On considering Fig. 14b, for flexible workplace, the RMSE of ORS-TSA-DNN at 75% learning percentage is 11.11, 8.57, 5.88, and 3.03% superior to DNN, SVM, KNN, and NN respectively. In Fig. 14c, for remote work, the RMSE of ORS-TSA-DNN at 55% learning percentage is 21.43, 18.52, 15.38, and 8.33% progressed than DNN, SVM, KNN, and NN respectively. Thus, the classifier analysis holds better outcomes with the ORS-TSA-DNN than the other algorithms for the HRM on work from home scenario method.

Fig. 14
figure 14

RMSE analysis of different classifier algorithms for the HRM on work from home scenario with respect to, a Employee wellbeing, b Flexible workplace, c Remote work, d Job loss, e Human capital, f Human resource development, g Leadership, h Performance, and i Communication”

Full size image

14.5 Overall prediction analysis

The overall prediction analysis of different heuristic-based and classifier algorithms for the proposed HRM method on work from home scenario during COVID pandemic situation is given in Tables 2 and 3 respectively. From Table 2, for leadership category, the RMSE of ORS-TSA-DNN is 3.54, 2.43, 1.27, and 4.57% higher than TSA-DNN, WOA-DNN, GWO-DNN, and PSO-DNN respectively. Similarly, for category performance, the RMSE of ORS-TSA-DNN is 3.56, 2.46, 1.28, and 4.61% superior to TSA-DNN, WOA-DNN, GWO-DNN, and PSO-DNN respectively. In Table 3, for communication category, the RMSE of ORS-TSA-DNN is 8.93, 7.84, 6.76, and 5.69% better than DNN, SVM, KNN, and NN respectively. Thus, the overall prediction analysis is better with the ORS-TSA-DNN than the other methods for the HRM on work from home during COVID pandemic situation.

Table 2 Overall prediction analysis of different heuristic-based algorithms for the HRM on work from home scenario
Full size table
Table 3 Overall prediction analysis of different classifier-based algorithms for the HRM on work from home scenario
Full size table

14.6 Evaluation of fivefold analysis

The evaluation of the different classifier and heuristic algorithms based on fivefold analysis for the HRM on work from home scenario is shown in Figs. 15 and 16. From Fig. 15a, for employee wellbeing, the RMSE of ORS-TSA-DNN by varying onefold values is obtained as 96.84, 94.54, 96, and 62.5% better than TSA-DNN, WOA-DNN, GWO-DNN, and PSO-DNN, respectively. Accordingly, Fig. 16c, for Remote Work, the RMSE of ORS-TSA-DNN by varying onefold values is secured 57.77, 65.45, 68.33, and 70.76% higher than DNN, SVM, KNN, and NN respectively. Therefore, it is proved that the evaluation of fivefold analyses attains better results with the ORS-TSA-DNN than the other different classifiers and heuristic algorithms for the HRM on work from home scenario method.

Fig. 15
figure 15

RMSE analysis of different heuristic algorithms for the HRM on work from home scenario with respect to, a Employee Wellbeing, b Flexible workplace, c Remote Work, d Job Loss, e Human Capital, f Human Resource Development, g Leadership, h Performance, and i Communication

Full size image
Fig. 16
figure 16

RMSE analysis of different classifiers for the HRM on work from home scenario with respect to, a Employee Wellbeing, b Flexible workplace, c Remote Work, d Job Loss, e Human Capital, f Human Resource Development, g Leadership, h Performance, and i Communication

Full size image

14.7 Comparitive analysis of the recent heuristic algorithms

The comparative analysis of the recent heuristic algorithms for the HRM on work from home scenario is shown in Fig. 17. From Fig. 17a, for employee wellbeing, the RMSE of ORS-TSA-DNN at learning percentage 75 is attained as 3.22, 12.90, and 6.45% lower than HRM, Tele Work, and Extant IB respectively. Accordingly, Fig. 16c, for Remote Work, the RMSE of ORS-TSA-DNN at learning percentage 45 is obtained 6.66, 20, and 13.33% minimized than HRM, Tele Work, and Extant IB respectively. Therefore, it is proved that the ORS-TSA-DNN is enhanced than the recent heuristic algorithms for the HRM on work from home scenario method.

Fig. 17
figure 17

RMSE analysis of recent heuristic algorithms for the HRM on work from home scenario with respect to, a Employee Wellbeing, b Flexible workplace, c Remote Work, d Job Loss, e Human Capital, f Human Resource Development, g Leadership, h Performance, and i Communication

Full size image

14.8 Discussion

The existing algorithms and classifiers contain certain demerits like, PSO takes a long time to deploy and gets low accuracy and the GWO, WOA, and TSA algorithms increase the error rate and provide a low exploitation phase and it does not support getting an optimal solution. The challenges of the different classifiers like SVM are data loss, high sensitivity, time constraints, and low accuracy, respectively, which also increase the error rate. For KNN, DNN, and NN the accuracy depends on the quality of data, complex sensor data and also it required high memory to store all the training data. HRM, Tele Work, and Extant IB provides better accomplishment to solve the optimization problems, in certain scenarios, it cannot be successful to get the optimum achievements because of the lower consistency, premature convergence, and local optimum points trapping. The proposed ORS-TSA-DNN model is more efficient than other conventional heuristic algorithms. This is due to the ability of ORS-TSA-DNN to produce the resultant outcomes with fewer errors and also contains fault tolerance capacity. It has a good exploitation capability, it obtained maximum accuracy, and the error rate is low. For these advantages, ORS-TSA-DNN holds better output than the other conventional algorithms and classifiers.

15 Conclusion

The purpose of this study was to look into HRD in IT businesses in Tamil Nadu in the case of a COVID pandemic, using a work-from-home scenario. "Employee wellbeing, flexible workplace, remote work, job loss, human capital, human resource development, leadership, performance, and communication" were among the nine areas examined here. A questionnaire was being prepared to address the issues that arise during the above-mentioned drives. The responses were then gathered from a variety of IT HR specialists in Tamil Nadu. These replies were used in the prediction phase, which was carried out with the proposed DNN that has been suggested. The ORS-TSA was used to optimize the weight of DNN in order to improve performance by reducing RMSE. From the analysis, the RMSE of ORS-TSA-DNN was 3.56, 2.46, 1.28, and 4.61% superior to TSA-DNN, WOA-DNN, GWO-DNN, and PSO-DNN respectively. Similarly, the RMSE of ORS-TSA-DNN was 8.93, 7.84, 6.76, and 5.69% better than DNN, SVM, KNN, and NN respectively. Thus, the overall prediction analysis is better with the ORS-TSA-DNN than the other methods for the HRM on work from home scenario during the COVID pandemic breakthrough. The practical implications for the HRM on work from home scenario are, AI technologies and tools play a key role in every aspect of the COVID-19. Deep learning models can help for real-time applications to predict old and new drugs or treatments that might treat COVID-19. The limitations of the research work is, to better investigate the role of communication with respect to different tasks and to the level of the related techno stress, it could be helpful to conduct a longitudinal study, in particular to understand the construction related to well-being that could depend on the particular experience of job or of the situation. Longitudinal analyses can be helpful in the understanding of the relation between communication and psycho-physical disorders, also observing stress fluctuations on a daily basis. Moreover, in order to better understand the active role of employees and to develop also positive and protective practices, future studies should consider the role of the individual self-efficacy related to the use of technology, which might impact on techno stress.