Movie Rating Prediction

Objective

Improve Accuracy For Predicting Movie Ratings using Machine Learning

Performed data collection. There are many datasets in Kaggle however many of the datsets were not very clean and many of the features we thought that would be important were not present. So we managed to combine three datasets for our project.
Performed data cleaning and preprocessing. Due to the diverse nature of movies we had to perform extensive data cleaning. Many of the fields in the features were catgorical and had many categories for example, genre, so we had to use many techniques to reduce the sparsity of the data.
Performed model traning using multiple regression models. Since the RMSE was quite high we also experimented with ensemble methods.

The best model was conculded to be Voting Regressor Ensemble method as it obtained a Test RMSE of 1.051.