- Prepare for the input data: In a csv file, each row is a time point, each column is a sensor measurement.
- Add a 'label' column.
- Define a time window you want to notify potential failures, such as 30 hours
Pass 3 arguments:
- path where training data is,
- sliding window size,
- names of the columns for feature extration,
- path where you want to save your trained model
data1:
python3 template.py /data/preventive_maintenance/train_turbofan/ 5 s1,s2,s3,s4,s5,s6,s7,s8,s9,s10,s11,s12,s13,s14,s15,s16,s17,s18,s19,s20,s21 /data/preventive_maintenance/model.pickle
python template.py ./train_turbofan/ 5 s1,s2,s3,s4,s5,s6,s7,s8,s9,s10,s11,s12,s13,s14,s15,s16,s17,s18,s19,s20,s21 ./turbofan_model.pickle
data2:
python3 template.py /data/preventive_maintenance/train_bearing/ 2560 Horizontal_acceleration,Vertical_acceleration /data/preventive_maintenance/model.pickle
python template.py ./train_bearing/ 2560 Horizontal_acceleration,Vertical_acceleration ./bearing_model.pickle
pass 2 arguments:
- path where you saved your trained model,
- path where testing data is
data1:
python3 test.py /data/preventive_maintenance/model.pickle /data/preventive_maintenance/test_turbofan.csv
python test.py ./turbofan_model.pickle ./test_turbofan.csv
data2:
python3 test.py /data/preventive_maintenance/model.pickle /data/preventive_maintenance/test_bearing.csv
python test.py ./bearing_model.pickle ./test_bearing.csv