TAD-FPMC

Train model

here are code used in train step.

1. random.h
2. matrix.h
3. tensor.h
4. pitf.h
5. train.cpp

use "g++ -o train train.cpp" to compile the code.

The train program will read train.txt and then generate 6 parameter file (l_p.txt, l_t.txt, l_u.txt, p_l.txt, t_l.txt, u_l.txt). The format of train.txt

user_id	time	current_POI_category	delta_T	next_POI_category
1	3	61	47.11	160
1	3	61	47.11	147
1	5	147	0.03	108

Predict

run predictCate.py to generate predict next_POI_category, and use cateprecision.py to get precision.