Remarks about data loading: 
The unzipped datafiles are bigger than 100mb therfore we can't push them to github. So everbody must unzip them themselfes afeter cloning the repo. The unzipped files are in the gitignore


Logs of all test runs:
18.10 22:05  / 8 features filterd out and mean of missing values with least square /  Categorical Accuracy = 0.733
19.10 13:12  / Filter feautures according to their absolut pearson cofficent > 0.1 / Categorical Accuracy = 0.706



Next steps to be done:
- implemnt logistic regression
- cross validation
- implement reglogistic regression