README

This repository contains the final group project for DSE 511 for Team Lancaster Barnstormers.

Team Members

• Tom Allemeier
• Inzamam Haque
• Verónica G. Melesse Vergara
• Harry Zhang

Project Goal

Deeper dive into the data from the individual projects by performing some statistics and machine learning to try and extract more patterns from the data.

Project Plan

• Prepare both oversampled and original data
• Train kNN, SVM, and RF repectively on two datasets and compare
• Write final report

Instructions of code

• Reading data by using

import prepare_data
X_origin, y_origin, X_over, y_over = prepare_data.get_data('path/to/your/database/')

• X contains features and y is the target variable (1: good loan, 0: bad loan)

• Running SVM classifier

import svm
clf0, X_test0, y_test0, clf1, X_test1, y_test1 = svm.run_svm(X_origin, y_origin, X_over, y_over)

• Running kNN classifier with the default k=1:

python3 knn.py

- Running kNN classifier with a different value of k:
```python

import knn
clf0, X_test0, y_test0, clf1, X_test1, y_test1 = knn.run_knn(X_origin, y_origin, X_over, y_over, 1)

• Running RF classifier

import RF
clf0, X_test0, y_test0, clf1, X_test1, y_test1 = RF.run_randomforest(X_origin, y_origin, X_over, y_over)

• Running the overall excecution script in terminal This prints out performance of all the classifiers and saves figures for ROC curve.

python excecution.py