Credit_Risk_Analysis
Overview of the analysis
Credit risk is an inherently unbalanced classification problem, as good loans easily outnumber risky loans.
We are going to use a number of different techniquest to train and evaluate models with unbalanced data.
Results
The below chart shows the confusion matrix and the related metrics
Below is the the chart of the metrics for all the models
- We see that all the models score high on low_risk prediction (high precision)
- We see that all the models peform poorly on high_risk prediction (low precision)
- The ensemble models work better on high_risk prediction
Summary
- The best model overall on both high and low risk prediction is Easy Ensemble Classifier