coronado212 / credit_worthiness_supervised_ml_12

Credit Risk Analysis Report

Overview of the Analysis

Credit risk poses a classification problem that’s inherently imbalanced. This is because healthy loans easily outnumber risky loans. Knowing that information, I leveraged various techniques to train and evaluate models with imbalanced classes in order to accurately analyze credit risk.

I leveraged a dataset of historical lending activity from a peer-to-peer lending services company, with the end goal being able to identify the creditworthiness of borrowers.

In order to have a full analysis, I broke the steps down into the following:

• Split the Data into Training and Testing Sets

• Create a Logistic Regression Model with the Original Data

• Predict a Logistic Regression Model with Resampled Training Data

• Write a Credit Risk Analysis Report

The machine learning model and techniques used for the analysis are the following:

• Logistic regression model - used to compare two versions of the dataset

  • 1st version - original dataset
  • 2nd version - resampled data using the RandomOverSampler module from the imbalanced-learn library

• For both versions, I did the following:

  • Counted the target classes
  • Trained a logistic regression classifier
  • Calculated the balanced accuracy score
  • Generated a confusion matrix
  • Generated a classification report

Results

• Machine Learning Model - Original Dataset:

• Machine Learning Model - Resampled Dataset:

Summary

• The logistic regression model fit with the oversampled data has a higher accuracy (99%) compared to the original data(95%). More importantly, it has a better specificity, which means less mistakes on false negatives and more accurate on the true negatives. From a business standpoint, this is extremely valuable, as you will lose more money on the riskier loans, so there's a higher need for accuracy there. Focus on the 1's!

---

Technologies

This project leverages python 3.7 with the following packages:

• imbalance-learn - A python package offering a number of re-sampling techniques

• PyDotPlus - Provides a Python Interface to Graphviz's Dotlanguage

---

Installation Guide

Install imbalance-learn by running the following command:

conda install -c conda-forge imbalanced-learn

Install PyDotPlus by running the following command:

conda install -c conda-forge pydotplus

---

Contributors

Brought to you by Edgar Coronado

---

License

MIT