Prediction of Type 2 Diabetes Mellitus Using Machine Learning Techniques

Among those critical diseases, Diabetes Mellitus is one of the

chronic diseases which affect human well-being at a young stage. The

chronic metabolic disorder diabetes mellitus is a rapidly growing global

challenge imposing massive socio-economic and health hazards. It has been

estimated that by the year 2020 there are nearly 285 million people (close to

6.4% of the adult age group) who are affected by this disease. This number

has been estimated to rise to 430 million with no better control or treatment

available. This rise in the rates in developing countries adopts the trend

changes in urbanization and lifestyle, which includes a "western-style" diet

also. This is due to the awareness being low . An aging population and

obesity constitute are the primary reasons for the rise.

In order to examine the high-risk population group of Diabetes

Mellitus (DM), modern information technology has to be used. Data mining

also called Knowledge Discovery in Databases (KDD) is defined to be the

computational process of finding the patterns in massive datasets that include

techniques intersecting Artificial Intelligence, Machine Learning, Statistics,

and Database Systems. The important objectives of these techniques include

Pattern Identification, Prediction, Association, and Clustering. Data

mining consists of a set of steps executed either automatically or semi-automatically

for extracting and finding intriguing, unknown, unseen features

from a paramount volume of data. The superior quality of data and the rightly

used technique are the two important concepts of data mining principle.

Several computational approaches have been designed for the

classification of diabetes occurs in humans. The usage of Machine Learning

in the medical information system has been found to be advantageous since it

improves the diagnostic accuracy, minimizes the expenditure, and also

increases the number of treatments that have been successful for diabetes

mellitus . For the automation of the overall process of diabetes prediction

and severity estimation, a diabetic database is required. This archive of the

diabetic database aids in identifying the effect of diabetes on different human

organs.