Simple Implementation of Gradient Boosted Trees

A Jupyter Notebook explaining the Gradient Tree Boosting algorithm [German]

This project compares multiple bagging and boosting methods for anomaly detection for the Gecco challenge.

Gained insights into the New York City Airbnb rental properties and discovered the trends in price and customer satisfaction level. Also discovered the kind of rentals receive what type of satisfaction level and predicted the likely satisfaction level of the new rentals leveraging advanced machine learning clustering algorithms such as k-means a…

Different feature selection methods like SVD, SelectKBest, RFE are used to choose best features, then different machine learning algorithms like Random Forest, Gradient Boosting Tree, XGBoost together with GridsearchCV etc are applied and compared to choose the good model which is the best fit for the dataset.

Few simple implementations of common ML algorithms from scratch (numpy)

A Spark application for weather forecasting using ensemble of tree-based models, trained on long-term historical data.

This is team 2's work for Project 2.

Using XGBoost , Gradient Boosted Trees to perform advanced regression and classification on structured tabular data

Developed a model/Spark ML pipeline stream to identify potential customers that may purchase top up services in the future.

In this repository, I implemented Gradient Boosting Trees using XGBoost to predict customer churn. The "Churn-modeling" dataset was downloaded from Kaggle.

Basic implementation of Gradient Boosting Trees

Worked on three use cases- Churn data analysis, Movie recommendation engine and Intrusion detection system.

Fraud detection on mobile banking transactions

A Python Package for a Sparse Additive Boosting Regressor

sentiment analysis using the movie reviews from the imdb database

In this project I wanted to predict attrition based on employee data. The data is an artificial dataset from IBM data scientists. It contains data for 1470 employees. Te dataset contains the following information per employee: