What is GiveMeSomeCredit?
GiveMeSomeCredit is a competition on Kaggle where participants aim to improve the state of the art in credit scoring. The task involves predicting the probability that an individual will experience financial distress in the next two years. In this Kaggle competition, participants use their data science and machine learning skills to develop models that can enhance the accuracy of credit scoring systems. By predicting the likelihood of financial distress within a two-year timeframe, the goal is to contribute advancements to the field, ultimately aiding in more informed decision-making processes for lending institutions.
What am I doing?
I conducted data analysis to understand the data's structure. Following that, I processed the data by removing outliers, eliminating duplicates, and verifying if any issues persisted after processing. Subsequently, I performed modeling using RandomForest, XGBoost, and LGBM. I consolidated all these results with an Ensemble Classifier, utilizing the soft method to observe probability outcomes. Once I ensured everything was correct, I submitted the results to Kaggle and obtained a Private score of 0.86634 and a public score of 0.86009.