Mutote, Michael, 22202956
Gattousi, Fadi, 22211572

Diamond Price Prediction Project

Recommendation Systems Project 2 https://mygit.th-deg.de/mm13956/ws-23-sas-02

Task Distribution

Project Overview

This project aims to predict diamond prices using four different regression models available in scikit-learn. It includes a Python script (Main.py) with a graphical user interface (GUI) for visualizing and interacting with the data and predictions. The dataset (diamonds.csv) consists of various attributes of diamonds. A Jupyter notebook (training.ipynb) is provided for data preprocessing, model training, and evaluation.

Setup and Installation

To set up this project, you need to install the required Python packages. You can do this by running the following command in your terminal:

pip install -r requirements.txt

Usage

1. Running the Application: Execute Main.py to launch the application. This script uses the cleaned data from diamonds.csv and a pre-trained Random Forest Regressor model to predict diamond prices.
2. Interacting with the GUI:
   • The GUI allows you to visualize data and predict diamond prices. It consists of two Tabs:
   • Main Tab: Enter values for the diamond attributes and click on the Calculate Price button to predict the price of the diamond. The predicted price will be displayed in the Predicted Price field at the bottom. In addition, the main window displays a scatter plot of the carat weight vs. the price of the diamonds in the dataset.
   • Advanced Tab: This tab allows you to visualize the relationship between the price and the other attributes of the diamonds. Select an attribute from the dropdown menu and click on the Plot button to display the scatter plot. Also you can select the model you want to use for prediction from the dropdown menu and choose which attributes you want to use for prediction. The unselected attributes's fields will be disabled in the main tab. The Calculate Price button will be disabled until the model is re-trained.

Model Selection

The following models are available for prediction:

• Linear Regression: This model assumes a linear relationship between the dependent and independent variables. We used the LinearRegression class from scikit-learn to train this model. The model achieved an R2 score of 0.88 and a mean squared error of 1,896,296.20
• XGBoost Regressor: This model is an ensemble of decision trees. We used the XGBRegressor class from scikit-learn to train this model. The model achieved an R2 score of 0.98 and a mean squared error of 303,969.39
• Multi-layer Perceptron Regressor: This model is a feedforward neural network with multiple hidden layers. We used the MLPRegressor class from scikit-learn to train this model. The model achieved an R2 score of 0.91 and a mean squared error of 31,366,191.38 and takes the longest time to train.
• Random Forest Regressor: This model is an ensemble of decision trees. We used the RandomForestRegressor class from scikit-learn to train this model. The model achieved an R2 score of 0.98 and a mean squared error of 292,726.76 Making it the best model for this dataset.

Project Structure

• diamonds.csv: Dataset containing diamond attributes.
• Main.py: Main Python script with GUI for prediction and visualization.
• requirements.txt: List of Python packages required for the project.
• training.ipynb: Jupyter notebook for data preprocessing and analysis.
• src/: Folder containing images used in this README.
• models/: Folder containing the trained models.

References

• Diamonds Dataset link from Kaggle