@ Author - Sandip Dutta

This repository is for storing code for Machine Learning Codes:

• Model with Outliers - Source Code for Medium Article [link to article]
  • Contains Jupyter notebook for Robust Estimators, that is ML models robust to outliers
    • Theil Sen Estimator
    • Huber Regression
• qsar_oral_toxicity - To predict the toxicity of oral compounds [link to dataset]
  • Fixed Imbalance using SMOTETomek Technique
  • Used Logistic Regression to predict the toxicity. Achieved a f score of 0.93
• Churn_Modelling - To predict if customer will churn (that is leave the comapny or not)
  • Cleaned the datasets and Imputed values
  • Applied various models and predicted the score.
  • Implemented Neural Network model using Keras for prediction.
  • Final accuracy came out to be about 89%
• Classification - Contains file for classification tasks.
  • Applied models on Diabetes Dataset and Loan prediction dataset.
  • Reached about 80% accuracy.
• Dimensionality Reduction - For Understanding Data Compression techniques
  • Short file for Understanding PCA and implementation
• Parkinsons - Parkinsons Disease Prediction
  • Prediction of Parkinsons Disease using XgBoost
  • Achieved an accuracy of over 95%
• SAT_Score - Prediction of SAT Score of Students
  • Regression model for predicting the Scores of Students