The purpose of this project is to predict drug users whose adverse drug reaction will lead to fatality. The sample data used for this predictive modelling was obtained from the United States of America federal drug agency database. The database contains reports of drug adverse events which occured in various countries of the world. A total of 92130 drug adverse reports were extracted from the database with selected 52 attributes of each reported events.

1. The most important features of samples which had fatal drug adverse effect
2. The Predictive model which gives the best performance in prediction of fatal drug adverse reaction

The model pipeline was developed using the following stages of model development

1. Data Import and Quality control
2. Descriptive statistics
3. Exploratory data analysis and visualisation
4. Feature engineering and extraction
5. Data selection and scaling
6. Predictive model building and evaluation
7. Model hyper parameter tuning
8. Final model and evaluation
9. Model deployment 10.Project summary and report