Flexible tool for bias detection, visualization, and mitigation

A scalable, explainable Java Naive Bayes Classifier that works either in memory or on persistent fast key-value store (MapDB, RocksDB or LevelDB)

Robust regression algorithm that can be used for explaining black box models (R implementation)

Robust regression algorithm that can be used for explaining black box models (Python implementation)

Designed a Machine Learning model which takes newsgroup dataset and performs binary classification to predict if a given document has Atheistic or Christian sentiment. Used LIME library and PySpark. Performed feature selection to improve classifier’s performance.