This script implements multiple linear regression using a closed-form solution to find optimal fit. User can analyze a dataset of their choice by providing:

1. relative path to their CSV file
2. name of the target column

If these are not provided, a demo dataset included with the script is used. For datasets with two or less independent variables, there is an option to visualize the fit of the regression model.

Python version: 3.10.8

External packages:

1. pandas
2. numpy
3. matplotlib

References

1. Closed form solution for linear regression
2. Simple Linear Regression: A Practical Implementation in Python
3. Argparse documentation

• Create a virtual environment and install the dependencies specified in requirements.txt
• Navigate to the Linear_Regression folder in the terminal.
• Understand the usage of the script:

    python linear_regression.py -h

• For the demo version, run the script without any arguments: (Demo dataset is self-generated using the random module from numpy package)

    python linear_regression.py

• To provide your own dataset (as a CSV file):

    python linear_regression.py <PATH_TO_CSV_FILE> <TARGET_NAME>

• Visualization can also be skipped: