DIAGNOSIS OF DIABETES USING INTELLIGIBLE MACHINE LEARNING IN A POPULATION

Abstract

Introduction: In the recent past, to attain decision-making patterns and interesting decisions, in almost every field, demand for machine learning has increased. This has also been applied to health data where effective analysis uses different techniques to assess the data. Presently, health data is accurate, crucial, and sensitive requiring accurate analysis, where results achieved are vital and of prime importance. Machine learning has increased sensitivity, role, interest, and data analysis.

Objectives: The present study was conducted to predict and analyze diabetes using learning algorithms on the diabetes data and to comparatively analyze the algorithms.

Methods: In the present study, for preprocessing the dataset median method was used. Following the preprocessing, 10 different algorithms from machine learning were used and applied to the dataset of diabetes in the present study.

Results: The present study used a diabetes dataset having 8 features or symptoms to predict diabetes. Different machine learning results of the mechanisms were analyzed and compared, to attain a better classification technique. In researches conducted in the future, the data from the present study can be utilized.

Conclusion: The present study concludes that better detection results are seen with linear support vector machines compared to other machine learning processes.