Analysis and Detection of Diabetes Using Data Mining Techniques—A Big Data Application in Health Care

Abstract

In digitized world, data is growing exponentially and Big Data Analytics is an emerging trend and a dominant research field. Data mining techniques play an energetic role in the application of Big Data in healthcare sector. Data mining algorithms give an exposure to analyse, detect and predict the presence of disease and help doctors in decision-making by early detection and right management. The main objective of data mining techniques in healthcare systems is to design an automated tool which diagnoses the medical data and intimates the patients and doctors about the intensity of the disease and the type of treatment to be best practiced based on the symptoms, patient record and treatment history. This paper emphasises on diabetes medical data where classification and clustering algorithms are implemented and the efficiency of the same is examined.

Appendices

Appendix 1

Algorithm: Gaussian Naïve Bayes

Input: Dataset

Output: Classification into different categories

Algorithm Steps:

Step 1. Segment the data by the class.

Step 2. Calculate the probability of each of the class.

$$ Class\,Probabilit = \frac{Class\,Count}{Total\,Count} $$

Step 3. Find the average and variance of individual attribute x belonging to a class c.

• a. Let $_{{}} \mu_{x}$ be the average of the attribute values in x allied with class c.

• b. Let $\sigma_{x}^{2}$ represent the variance of the attribute values in x related with class c.

Step 4. The probability distribution is computed by

$$ {\text{p}}\left( {{\text{x}} = {\text{v|c}}} \right) = \frac{1}{{\sqrt {2\pi \sigma_{x}^{2} } }}e^{{\frac{{ - \left( {v - \mu_{x} } \right)^{2} }}{{2\sigma_{x}^{2} }}}} $$

Step 5. Calculate the probability of the attribute x

$$ P(x_{1} ,x_{2 \ldots \ldots \ldots } x_{n} |c) = \mathop \prod \limits_{i} P (x_{i} |c) $$

Naïve Bayes Classifier = argmax P(c) $\mathop \prod \limits_{i} P (x_{i} |c)$

Step 6. End.

Appendix 2

Algorithm: OPTICS

Input: Dataset

Output: Clusters

Algorithm Steps:

Step 1. Initially ε and MinPts need to be specified.

Step 2. All the data points in the dataset are marked as unprocessed.

Step 3. Neighbours are found for each point p which is unprocessed.

Step 4. Now mark the point as processed.

Step 5. Initialize the core distance for the data point p.

Step 6. Create an Order file and append point p to the file.

Step 7. If core distance initialization is unsuccessful, return back to Step 3 otherwise go to Step 8.

Step 8. Calculate the reachability distance for each of the neighbours and update the order seed with the reference of new values.

Step 9. Find the neighbours for each data point in the order seed and update the point as processed.

Step 10. Fix the core distance of the point and append to the order file.

Step 11. If undefined core distance exists, go back to Step 9, else continue with Step 12.

Step 12. Repeat Step 8 until there is no change in the order seed.

Step 13. End.

Appendix 3

Algorithm: BIRCH

Input: Dataset

Output: Clusters

Algorithm steps:

Step 1. Set an initial threshold value and insert data points to the CF tree w.r.t the Insertion algorithm.

Step 2. Increase the threshold value if the size of the tree exceeds the memory limit assigned to it.

Step 3. Reconstruct the partially built tree according to the newly set threshold values and memory limit.

Step 4. Repeat Step 1 to Step 3 until all the data objects are scanned and form a complete tree.

Step 5. Smaller CF trees are built by varying the threshold values and eliminating the Outliers.

Step 6. Considering the leaf entities of the CF tree, the clustering quality is improved by applying the global clustering algorithm.

Step 7. Redistribution of data objects and labelling each point in the completely built CF tree.