In this project, your task is to perform feature engineering and feature scaling $($including transformations like logarithmic scaling$)$ on our dataset to enhance the performance of a linear regression model. You have complete freedom to decide how many features to include, which features to use, and in what formats to present them for the linear regression model.

Please ensure the following:

You are required to manually create at least five new features that are different from those used in this lecture.

If you decide to apply feature scaling $($e$.$g$.,$ robust scaling$),$ be careful to avoid any data leakage.

Feel free to explore polynomial features and utilize feature selection techniques.

Use $5-$fold cross$-$validation with a random seed of $42,$ as shown in the notebook, to run and validate a linear regression model. Report its performance across all five folds, along with the averages. The performance metrics should include R$-$squared, RMSE, and MAPE.

Your grade will mainly depend on the model's performance, especially the average R$-$squared value. This value will then be scaled to determine the final score for the project. Your model should at least outperform the benchmark of $0\mathrm{.}7338,$ but ideally, it should exceed $0\mathrm{.}8095,$ which is the highest R$-$squared value in this notebook.

Since the dataset split and model are predetermined, the results and overall performance will rely entirely on the feature engineering and scaling techniques you implement. Ultimately, Ultimately, this project aims to enhance your understanding of the crucial role these tasks play in model performance.