SeniorProject

Cal Poly Senior Project 2018-2019

Results of running oversampling, undersampling, and SMOTE sampling methods on various datasets using WEKA api.

Guide to Files

Each dataset has a folder. Within the folder, you will find the following files/directories:

1. <dataset name>Undersampling - contains 100 result files from the dataset being undersampled

2. <dataset name>Oversampling - contains 100 result files from the dataset being oversampled

3. <dataset name>SMOTE - contains 100 result files from the dataset being sampled with SMOTE

4. <dataset name>_A.arff - the unsampled (no sample) file (for training)

5. <dataset name>_B.arff - the unsampled (no sample) file (for testing)

6. <dataset name>.csv - the original dataset prior to sampling, column removal, and splitting into A and B.

7. <dataset name>_A.csv - the dataset prior to sampling and non-featured column removal, but post splitting into A and B (for training)

8. <dataset name>_B.csv - the dataset prior to sampling and non-featured column removal, but post splitting into A and B (for testing)

RQ3 Details

This repository contains the following datasets for RQ3 Undersampling and Oversampling:

1. Ant
2. Camel
3. Ivy
4. Jedit
5. Keymind-A
6. Keymind-B
7. Log4j
8. Lucene
9. Poi
10. Prop (only Undersampling 1%)
11. Synapse
12. Velocity
13. Xalan
14. Xerces