Personal lending is growing faster than credit card, auto, mortgage, and even student debt. With such incredible growth, FinTech firms are storming ahead of traditional loan processes. By using the latest machine learning techniques, these FinTech firms can continuously analyze large amounts of data and predict trends to optimize lending.
In this project, Python has been used to build and evaluate several machine learning models to predict credit risk. Being able to predict credit risk with machine learning algorithms can help banks and financial institutions predict anomalies, reduce risk cases, monitor portfolios, and provide recommendations on what to do in cases of fraud.
Python, numpy, scikit-learn, tensorflow, logistic regression, decision tree, random forest, support vector machine algorithms, ensemble and resampling techniques
Credit risk is an inherently unbalanced classification problem, as the number of good loans easily outnumber the number of risky loans. Therefore, we need to employ different techniques to train and evaluate models with unbalanced classes. The following techniques have been used in this analysis to access the credit risk, using data from LendingClub; a peer-to-peer lending services company.
The accuracy score is 0.63 which means that the model is correct 63% of the time. From the imbalanced classification report, the precision is 1 for low-risk(majority class) and 0.01 for high-risk(minority class) loans which is not a good value to rely upon. The recall for high-risk loans is slightly better than the low-risk loans but F1 score is very poor for the high-risk loans. Hence, this model is not a good predictor to assess the credit risk.
The accuracy score is 0.66 which means that the model is correct 66% of the time. From the imbalanced classification report, the precision is 1 for low-risk(majority class) and 0.01 for high-risk(minority class) loans which is not a good value to rely upon. The recall for high-risk loans is nearly equal to the low-risk loans but F1 score is very poor for the high-risk loans. Hence, this model is also not a good predictor to assess the credit risk.
The accuracy score is 0.53 which means that the model is correct 53% of the time. From the imbalanced classification report, the precision is 1 for low-risk(majority class) and 0.01 for high-risk(minority class) loans which is not a good value to rely upon. The recall for high-risk loans is greater than the low-risk loans but F1 score is very poor for the high-risk loans. Hence, this model is also not a good predictor to assess the credit risk.
The accuracy score is 0.65 which means that the model is correct 65% of the time. From the imbalanced classification report, the precision is 1 for low-risk(majority class) and 0.01 for high-risk(minority class) loans which is not a good value to rely upon. The recall for high-risk loans is greater than the low-risk loans but F1 score is still very poor for the high-risk loans. Hence, this model is also not a good predictor to assess the credit risk. But this model is better in terms of recall as compared to the above models. This model may be used if we give more weightage to the recall for high-risk loans.
The accuracy score is 0.77 which means that the model is correct 77% of the time. From the imbalanced classification report, the precision is 1 for low-risk(majority class) and 0.03 for high-risk(minority class) loans which is not a good value to rely upon. The recall for high-risk loans is less than the low-risk loans but F1 score is still very poor for the high-risk loans. Hence, this model is also not a good predictor to assess the credit risk. This model can rather be used to access low-risk loans better than the high-risk loans.
The accuracy score is 0.93 which means that the model is correct 93% of the time. From the imbalanced classification report, the precision is 1 for low-risk(majority class) and 0.09 for high-risk(minority class) loans which is not a good value to rely upon. The recall for both high-risk loans and the low-risk loans is very high. Although F1 score is 0.16, but it is better than all the above models for the high-risk loans. Hence, this model may be used to assess the credit risk bit it is still not a good predictor.