This repository provides the code of our Project "Ground recognition via machine Learning". The code is split into the parts of data preprocessing, feature engineering and machine learning.
- In the preprocessing the samples are cropped to the right period and processed with a suspension coefficient. Furthermore the data gets filtered by a Bandpass-Filter and outliers are removed via distribution-based outlier detection.
- During the feature engineering the samples get split into 10s snippets. On each of them we calculate statistical and frequency features.
- In the last step the features are used to train different machine learning models (KNN, DT & SVM) and optimize them for validation- and test-data.
- Numpy
- Pandas
- Matplotlib
- Scikit-Learn
- Scipy