classification_and_regression

General info

Repository for second project in FYS-STK 3155. All methods are found in code/. Note that the plots in visuals/ may differ from those in report.

How to run code

• install pipenv on your system
• clone the repository
• in correct folder, type:

install pipenv

• enter shell:

pipenv shell

• run code file as normal

Table of contents

• Visuals
• Classes
• Example runs
• Test runs
• Benchmark results
• Report

Example use

To run SGD, Logistic Regression and Neural Network methods, run example files in example_runs/.

stochastic_run.py:

Study hyperparameters and find best set for SGD

logistic_run.py:

Find best set of hyperparameters for logistic regression

neural_network_run:

Find best set of hyperparameters for the Feed Forward Neural Network. for classification and comparing with sklearn and OLS.

neural_network_regression:

Find best set of hyperparameters for the Feed Forward Neural Network. for regression and comparing with sklearn.

Test of codes

to test the methods write

pytest -v

while in code/tests or run the program normally. this also serves as benchmarks to check that each part of the code is running as expected.