In this project, I have utilized survival analysis models to see how the likelihood of the customer churn changes over time and to calculate customer LTV. I have also implemented the Random Forest model to predict if a customer is going to churn and deployed a model using the flask web app.

Variable Importance Plots (VIPs)

A general framework for constructing partial dependence (i.e., marginal effect) plots from various types machine learning models in R.

Python package to visualize and cluster partial dependence.

Material from "Random Forests and Gradient Boosting Machines in R" presented at Machine Learning Day '18

🎯📈 Sequantial and model-based optimization

Data Mining Final Project

Predict churning or not from the real-world data of a ridesharing app

Complex odor analysis and interpretation

Predicting product recommendation score using the data available on the website of the client

The goal of SHAP is to explain the prediction of an instance x by computing the contribution of each feature to the prediction. The SHAP explanation method computes Shapley values from coalitional game theory. The feature values of a data instance act as players in a coalition.

This project aims to study the influence factors of international students' mobility with the case of international students from B&R countries studying in China.

Kaggle kernels and the respective implementations of ML procedures.

This project contains the data, code and results used in the paper title "On the relationship of novelty and value in digitalization patents: A machine learning approach".

Data Science Case Study

Partial dependence plot tool

Robustness of Global Feature Effect Explanations (ECML PKDD 2024)

Meta-analysis of learning and memory in PTSD

Trained a classifier by using labeled data and oversampling and undersampling techniques to predict if a borrower will default on a loan. The model is intended to be used as a reference tool to help investors make informed decisions about lending to potential borrowers based on their ability to repay. The purpose is to lower risk & maximize profit.

Individual Conditional Expectation (ICE) plots display one line per instance that shows how the instance's prediction changes when a feature changes. The Partial Dependence Plot (PDP) for the average effect of a feature is a global method because it does not focus on specific instances, but on an overall average.