This project aims to use the regularised regression models such as lasso regression, ridge regression, and elastic nets to identify the variables that predict whether a basketball player will play more than five seasons in the NBA. See details in Project memo.pdf

Building a predictive model for the given data set to test it on test data set

The purpose of the present project is to create a recommendation system for predicting the rating of movies.

Recherche de composante(s) explicative(s) par pénalisation (RCEP)

regDIF: Regularized Differential Item Functioning

Predictive Model on House Price

R Package: Regularized Principal Component Analysis for Spatial Data

Imbalanced classification with loan clients dataset.

R code for figures, simulations and application in my Masters thesis and corresponding conference paper on the LASSO and other shrinkage methods

Statistical Learning application of different machine learning algorithms on dataset and their implementation in R covering model selection techniques, shrinkage Methods and Regularization techniques, different non linear models, re-sampling methods and Cross Validation, boosting, trees, supervised learning & Unsupervised learning

This is one of my final projects for the HarvardX Data Science Professional Certificate Program. As the title suggests, it is on the GroupLense database colloquially known as MovieLens. The goal of the project is to predict ratings with a RMSE below .86490. I was able to surpass the goal with 3 different models. Happy reading!

R1magic: Compressive Sampling: Sparse signal recovery utilities

Support vector machines flexible framework

Multi-source sparse Tweedie modelling

Compared the mean square error and R-Square value using linear model, Lasso and ridge regression

Repository for my Master's thesis comparing volatility models.

Jackstraw Weighted Shrinkage Methods

lava plugin allowing regularization terms

Base R Implementation of Logistic Regression from Scratch with Regularization, Laplace Approximation and more