anjapago / logisiticregressionclassifier Goto Github PK

For explanations, please view the report pdf, or the python notebook.

This project contains three files that have the logistic regression implementation.

1. "logisticregression.py":
This can be run from the terminal with python 3 and the desired data file. For example to run this, the command would be:
python3.6 logisticregression.py "data.mat"
(using one of the .mat files in the data folder)

This file will run both the training and testing algorithms, and output results of the accuracy. Also it will display loglikelihoods as it iterates.

2. "Logistic-Regression-Classifier.ipynb"
This can be opened as a jupyter notebook and run step by step. Running each box in order will result in running the full logistic regression with gradient descent. It requires python 3 as the kernel, and all the necessary imports. It requires the data files and the classifier.py file.

3. "classifier.py"
This is a simplified version of "logisticregression.py" containing code meant to be used from the jupyter notebook.

Note: first unzip and move out the data files.