ft_linear_regression

Project for training linear models (with multiple features) and predict by trained model

Install and run

> ./install.sh
> ./train.py args...
> ./predict.py args_for_predict...

OR

(venv) > pip3 install -r requirements.txt
(venv) > python3 train.py args...
(venv) > python3 predict.py args_for_predict...

Usage

train.py

• path/to/file.csv - first argument, path to dataset file in csv format
• -p - flag to show precision value on every iteration
• -v - flag to vizualize trained model with dataset and cost function iterations
• --alpha=0.01 - get learning rate value to 0.01
• --iter=1000 - get maximum iterations value to 1000

predict.py

This script run in format

> ./predict.py arg1 arg2 arg3 arg4...

arg1 arg2 arg3 arg4... - are real numbers. They are feature set