Data of the study "A method to evaluate task-specific importance of spatio-temporal units based on explainable artificial intelligence"

Data (numbered one to four) are saved as the byte format and can be read by the pickle library of python.1. The file "holiday_2_test_2016" is the organized taxi origin point distribution dataset collected in the whole year of 2016. This dataset is used as the validation set of model training. Each input data are labeled by the temporal information (weekday or weekend/holiday).2. The file "holiday_2_train_2016" is the organized taxi origin point distribution dataset collected in the whole year of 2016. This dataset is used as the training set of model training. Each input data are labeled by the temporal information (weekday or weekend/holiday).3. The file "holiday_2_test_2017" is the organized taxi origin point distribution dataset collected in three months (from January to March) of 2017. This dataset is used as the test set. Each input data are labeled by the temporal information (weekday or weekend/holiday). 4. The file "totalmean" is the mean value distribution (30*30*48) of the taxi origin point volume. This dataset is used as the input data of the clustering experiment.5. The file "POIStatistics.csv" is the contribution value difference and POI counts of units. This dataset is used as the input data of the regression experiment.<br>