Data transformation and classificiation methods for J-PET tomography.

Requirements: python3.6 and the required packages
Input format: CSV
Input columns: Described in the function dataFrameNames() at /RepairingData/repairingData.py script
Output format: CSV
Output columns: same as above
Purpose: Repairing wrong class labels and shuffle the order of particles.
Usage: ./RepairingData/repairingData.py 'path_to_output_data' 'data_part'
PS: Input data folder and filename set as '/mnt/opt/groups/jpet/NEMA_Image_Quality/3000s/' and 'NEMA_IQ_384str_N0_1000_COINCIDENCES_'. You can change it inside the script.

Requirements: python3.6 and the required packages
Input format: CSV
Input columns: Discribed in the function dataFrameNames() at /TransformingData/transformingData.py script
Output format: binary file
Output columns: same as above + result of the function featureEngineering() at /TransformingData/transformingData.py script
Purpose: Feature engineering - add extra features (listed in the variable newAttributes in /TransformingData/transformingData.py script) to the original ones (listed in the variable originalAttributes in /TransformingData/transformingData.py script). After the transformation new data (in pandas DataFrame format) is saved as binary file (serialization made using the Pickle package).
Usage: ./TransformingData/transformingData.py 'path_to_data' 'data_part'
PS: Filename set as 'NEMA_IQ_384str_N0_1000_COINCIDENCES_' with the infixion 'REAIRED_' while loading the data (see the scirpt /TransformingData/transformingData.py). You can change it inside the script.
Output usage example: ./TransformingData/transformingDataExample.py 'path_to_the_transformed_file'