Diabetes prediction using machine learning

Hospitals contain large databases. We can use that data to discover new useful and potentially life saving knowledge. Here we use datamining especially to predict type 2 diabetes mellitus.Predicting the percentage of chance of occurrence of Diabetes mellitus type 2 with less time complexity and high accuracy.

Algorithms used

1. Random forest algorithm.
2. K-Means.
3. Logistic Regression.
4. Naive Bayes.
5. Support Vector Classifier.

concepts used

1. Recursive Feature Elimination (RFE).
2. Recursive Feature Elimination with Cross Validation (RFECV).
3. Priciple Component Analysis.
4. Grid Search.
5. Voting Classifier.

The idea behind using all these algorithms and different concepts is to draw out comparison between them. In the python file attached we neatly show the comparision graph for Logistic Regression, Random Forest and Naive Bayes Algorithm outputs with out RFE, With RFE and with RFECV.