+classifier ----+model: trained model saved ----+cnn.py: cnn code ----+dataloader.py: data loader from ../data ----+stackClassifier.py: stacked classifier of cnn+svm **executable** ----+svm.py: svm code ----+svm.script.py: svm test code +data ----+datfiles: 220 labeled data (raw data, group by vins and sorted) ----+datsample: samples from ./datafiles ----+sjtu:sjtu dataset --------+allmat: all map (in matrix on vin/daily basis) --------+Bus: -------------+img: all png files of map matrix -------------+raw: raw data --------+Taxi: -------------+img: all png files of map matrix -------------+raw: raw data --------+label.csv : labels ----+label.csv: label for 220 labeled data in datfiles +feature ----+map ---------+full: whole day matrix ---------+img: png file for map matrix ---------+part: map matrix of (0:0-6am,1:6-12am,2:12-18pm,3:18-24pm) ----+featureEngineering.py: extract features for 220 lableled data, output to feature_extracted.txt and ./map **executable** ----+mapSim.py: calculate map similarity ----+feature_extracted.txt: svm feature extracted +transfer: all code for transfer learning from sjtu dataset to 220 labeled dataset ----+cnn.py ----+crossvalidation.py: executable ----+dataLoader.py ----+labels.py: generate labels ----+processDataBus/Taxi.py: generate feature ----+rawdataloader.py: load 220 labeled raw data ----+transferacc.py: calculate transfer accuracy +readme