An ML powered model to predict games user ratings
This project is a machine learning model that predicts the user rating of a game based on the game's features. The model is trained on a dataset of 5,214 games and their features. The model is trained using various regression and classification algorithms. The model is deployed using streamlit.
The dataset used for this project is a portion of a Kaggle dataset. The dataset is provided by the FCIS ML Team as a part of the FCIS ML course project. The dataset contains 5,214 games and their features.
The dataset used for this project can be found in the repository here for regression and here for classification. The original dataset can be found here.
- Name: The name of the game
- Subtitle: The subtitle of the game
- Price: The price of the game
- Primary Genre: The primary genre of the game
- Genres: The genres of the game
- Languages: The languages of the game
- Size: The size of the game
- Original Release Date: The original release date of the game
- Current Version Release Date: The current version release date of the game
- Age Rating: The age rating of the game
- URL: The URL of the game
- Icon URL: The icon URL of the game
- Description: The description of the game
- Developer: The developer of the game
- User Rating Count: The user rating count of the game
- In-app Purchases: The in-app purchases of the game
- Rate: The rating of the game (Target for classification)
- Average User Rating: The average user rating of the game (Target for regression)
The project report can be found here. The report contains how the project was developed, what were our approaches, what features were used, how did we preprocess the data, the algorithms used, the results, and the conclusion.
| Model | Train MSE | Val MSE | Test MSE | Train R2 | Val R2 | Test R2 |
|---|---|---|---|---|---|---|
| XGBoost | 0.1876 | 0.3264 | 0.4539 | 0.56 | 0.27 | 0.15 |
| GradientBoosting | 0.1634 | 0.3209 | 0.4530 | 0.61 | 0.28 | 0.15 |
| PolynomialRegression | 0.2546 | 0.3384 | 0.4611 | 0.40 | 0.25 | 0.13 |
| ElasticNet | 0.2805 | 0.3439 | 0.4641 | 0.34 | 0.23 | 0.13 |
| Linear Regression | 0.2795 | 0.3439 | 0.4644 | 0.34 | 0.23 | 0.13 |
| CatBoost | 0.1644 | 0.3133 | 0.4675 | 0.61 | 0.30 | 0.12 |
| Model | Train Accuracy | Validation Accuracy | Test Accuracy |
|---|---|---|---|
| SVC | 73.21% | 68.12% | 66.31% |
| RandomForest | 74.48% | 66.29% | 65.85% |
| LogisticRegression | 71.11% | 66.48% | 65.23% |
| CatBoost | 74.71% | 68.48% | 63.38% |
- FCIS ML Team: Provided the dataset (Originally a portion from a Kaggle dataset) and the guidance while developing the project.