Historical transactions, delayed transactions, account identity retrieval have generally been a bottleneck to financial institutions all over the world. A lot of financial institutions do not have quick access to user information and general data making it difficult to process transactions leading to delayed settlements, not having a good estimate of customer’s spending power and inability to accurately identify customers. To solve this issue together my team and I came up with a web application that allows financial institutions to track their users information by visualizing it and also to predict if a user should be allowed a loan amount or not.
The app allows for the loan company to know whether a user would default on a loan being requested or not using machine learning algorithms with factors such as Income, employment status, loan amount, loan time period and gender of the user.
For modeling, we use Classification algorithms such as Logistic regression, Support vector machine (SVM) algorithm, Random Forest Classification, Xgboost, K-nearest neighbours’ algorithm, and Decision Tree Classification and after training the best accuracy of 67% was gotten using the support vector machine algorithm and Logistic regression.
Ingressive For Good: https://ingressive.org/