Negative-Unlabeled Tensor Factorization for Location\ Context Inference from Inaccurate Mobility Data (NUTF)

This repository includes a simple matlab simulation and C++ codes for NUTF

Usage

• matlab version: run simulateDemo.m in ./matlab/

• cpp version:

Install Eigen (http://eigen.tuxfamily.org/index.php?title=Main_Page#Download) and put it in ./eigen/

Compile training file:

make train

./train N T C r iter
e.g.:
./train test.txt 14 6 3 5 5

Compile simulation file:

make simulate

./simulate N T C r iter k
(N: number of users, T: number of time steps, C: number of locations, r: rank, iter: number of iterations, k: sparsity)
e.g.:
./simulate 10000 500 200 20 10 20