Github-link for project 3: Superconductivity
This is a github repository with various models (SVM, linear regression, adaboost and XGBoost) in order to predict the critical temperature of known superconductors.
Folders:
Data: contains the files from the Kaggle page as well as the data which scraped from the same sources as the paper.
Results: contains the csv files and figures from the different methods: SVM, Adaboost, Linear Regression, XGBoost
Report: contain the pdf version of the report for FYS-STK4155 project 3
Files:
BaysianOptimizer.py contains a class for maximising a models score function with respect to the models hyper-parameters.
SVM_regression.py contains functions for the SVM model.
adaboost_example.py contains script for an example run using the Adaboost.
adaboosting.py contains a class 'AdaBoost' for running the adaboost.
helper_functions.py contains functions for supporting the different classes, e.g importing data.
linear_regression.py contains script for running linear regression functions, OLS, Ridge and LASSO as well as using Gridsearch and Bayesiansearch for hyperparameters.
methods.py contains a class Regression which has various regression model functions within it.
scrape.py is a script for scaping the japanese website for the oxide superconductor data.
test_adaboost.py is a unit test for comparing own implementation with sklearns adaboost
test_bays.py is a script for testing the BayesianOptimizer