Even though this is an Analytics Vidhya competition, my goal in this project is not to compete or to construct the most accurate model but to demonstrate basic skills of tackling cleaned datasets that includes: handling missing values, exploratory analysis, feature engineering, building predictive model, tuning model parameters, and model evaluation, as well as gaining insights from data and model. The goal of this problem is to predict the status of loan approval of test data set as accurate as possible.
Conclusion- This is the end of the analysis, we started from data cleaning and processing, missing value imputation with mice package, then exploratory analysis and feature engineering, and finally model building and evaluation. What is more important, we gain some insights about loan approval from our analyis, described below.
->Applicants with credit history not passing guidelines mostly fails to get approved, probably because that they have a higher probability of not paying back.
->Most of the time, applicants with high income, loaning low amount is more likely to get approved, which makes sense, those applicants are more likely to pay back their loans.
->Having a strong coapplicant can be a plus to the probability of getting approve.
->Some basic characteristic such as gender and the status of marriage seems not to be taken into consideration by the company.