Prediction Analysis using Weighted Product Method to Compare Machine learning Algorithms for Diabetes Disease

Authors

  • Jerry Malapane Department of Electrical and Electronic Engineering, University of Johannesburg, Gauteng, South Africa
  • W. Doorsamy Institute for Intelligent Systems, University of Johannesburg, Gauteng, South Africa
  • B. S. Paul Institute for Intelligent Systems, University of Johannesburg, Gauteng, South Africa

Keywords:

MCDM, Diabetes, Machine learning, Dataset, WPM

Abstract

Diabetes is one of the well-known and most serious chronic diseases in the world, causing a person to suffer from a raised level of blood sugar due to body resistance to producing the essential volume of insulin. People suffering from diabetes may have complications such as cardiovascular diseases, blindness, and kidney diseases. Early prediction and diagnosis of diabetes can save many lives by alarming people to require medical attention. One possible solution for the diagnosis and prediction of diabetes disease is the use of machine learning approaches. In this paper, prediction analysis of diabetes disease is performed using various machine learning techniques and comparative analysis based on the Multi-Criteria Decision-Making method to vote for the best algorithm. Few Machine learning techniques such as Support Vector Machine, K-NN, Random Forest Classifier, Naïve Bayes, Extreme Gradient Boosting, Adaptive Boosting, Multilayer Perceptron, and hybridized K-mean Random Forest. In our study experiment, we used PIMA Indian Dataset retrieved from UCI Repository. The results show that all evaluation criteria are valid and calculated using Weighted Product Methods to provide the best algorithm. Our experimental results show that the Extreme Gradient Boosting technique achieved the highest ranking when relating to the other seven machine learning algorithms.

Downloads

Download data is not yet available.

Downloads

Published

15-09-2022

How to Cite

[1]
J. Malapane, W. Doorsamy, and B. S. Paul, “Prediction Analysis using Weighted Product Method to Compare Machine learning Algorithms for Diabetes Disease”, IJRESM, vol. 5, no. 9, pp. 49–53, Sep. 2022.

Issue

Section

Articles