A Comprehensive Survey of various classification techniques in Data Mining for Diabetes data

Abstract

Data mining techniquesare considered as promising techniques for data analysis and have been
adopted widely in the medical field due to their nature of efficient analysis of huge data. Huge amount of
medical data are being acquired and stored which needs be monitored in an efficient man
ner. However, these datasets are uncertain and dynamic in nature which raises complexity to maintain
smooth access of the data. Moreover, these datasets are used for analyzing the symptoms of the disease
which can be useful to provide efficient diagnosis. In order to present this analysis, machine learning based
schemes have been adopted widely via data mining techniques. According to the machine learning based
data mining, historical medical data is analyzed through their attributes and disease is predicted based on the
data of the current patient. These schemes are developed for various types of diseases such as ECG signal
classification, cardiovascular disease monitoring and diabetes disease etc. In this work, we focus on the
diabetes prediction and present a brief review study about machine learning and data mining based solutions
to predict the diabetes. The popular machine learning techniques include Bayesian, Random forest
algorithms, Artificial Neural network, SVM and Decision Tree etc.