Implementing ADMM for solving convex optimization problems such as Lasso, Ridge regression
ADMM works in two steps
- Divide
a. Break down original problem into small problems
b. Distribute these small problem to <N> processors / computing resources
c. Every processor solves the smaller problem - Conquer
a. Combine solution from <N> processors into one
We are using the Fraud detection dataset from Kaggle for comparing the models. Fraud Detection Dataset
We implemented ADMM based and regular Lasso and Ridge regression models using python. We made a comparative study of these models on the provided datasets.