imad280/Algeria-data-cup-challenges

This repository showcases our efforts and contributions to an exciting set of challenges proposed in a datathon where the goal was to develop robust models to model the provided datasets. # Algeria-data-cup-challenges This repository showcases our efforts and contributions to an exciting set of challenges proposed in a datathon where the goal was to develop robust models to model the provided datasets. Loan propensity prediction : The objective of this challenge was to predict a bank's client propensity for loans ( link to the competition ). To attempt that, we modeled the dataset in 4 models: Decision tree, Random Forest, XGBoost and Multi-Layer Perceptron (MLP). The different accuracies are in the notebook. Overdue deadlines prediction : The objective of this challenge was to predict the number of overdue deadlines for a bank ( link to the competition ). As a solution, we attempted 3 models: Linear Regression, Random Forest and XGBoost. The results of the modelisation will be found in the notebook. Purchase prediction : The goal of this challenge is to predict whether a client will make a purchase on a specific date ( link to the competition . The model that we constructed was by using XGBoost. The results can be found in the notebook. Clustering challenge : This challenge was an additional task in the competiton. It consisted in a client segmentation task. The methods we used were: DBSCAN and K-means. The results are on the correspending notebook.