Data Mining (Weka)

A medical dataset is provided on Blackboard under Assignment->Project 2->diabetic.csv. This dataset is for prediction of hospital readmission of diabetic patients. It has 80k samples and 50 features. The dataset contains 3 classes: no readmission (NO), readmission under 30 days (<30), and readmission after 30 days (>30). First, become familiar with data set, and research some background information on diabetic readmission. Then, review published research papers (minimum are 3!) on applying data mining to medical data and briefly discuss their approach. Make sure you cite your sources (example format see appendix). Then, go through the detailed process of data preparation. Apply all preprocessing and data reduction techniques you assume they are necessary. When the data set is with enough quality, apply several predictive based techniques (minimum are 3 techniques!) and create appropriate predictive models. Test your models so that you can compare data mining techniques. Comparison of data mining techniques (and obtained predictive models) with additional discussion and interpretation of results will be very important part of your report. You will finish your project 2 report using this dataset. A separate dataset ([login to view URL]) will be provided for additional evaluation. It contains 20k unlabeled samples. You will give prediction results using one of your best prediction models. We will compare your prediction results against the true label for additional unbiased evaluation of your project. Up to 3 points in the final grade of the project will depend upon your model’s performance with this dataset.