This repository includes the code of MetaMLAnn (see paper).
-
MetaMLAnn folder contains the core code for our model.
-
naive_model.py: Evaluate baseline models. It will print out the result for 4 baseline models under K fold. you can change it. parameters: '--KFold' Number of Fold. '--embedding' 0 to use stations feature only; 1 to use embedding; 2 to use both (default)
-
feature_extraction.py: Functions to extract embedding features/ subway line features
-
gen_data.py: construct species matrix (label) from files or by loading from src/tmp/pic_spec_vec.data
-
vectorize.py: exploration function for species distribuion similarity v.s. distance. Generate the matrix. Similar as part of function in the gen_data.py.