hxsylzpf / experimental-mf

cache-friendly multithread matrix factorization

Fast Matrix Factorization

Features

• Cache-friendly Multithread Matrix Factorization.
• Fast Multithread Stochastic Gradient Langevin Dynamics (SGLD) for Matrix Factorization.
• Fast Multithread Differentially Private Matrix Factorization.
• Matrix Factorization with Adaptive Regularizer.

Data

We support Google's Protobuf as input. Try data/getdata.cc to convert from userwise raw data to protobuf format:

./getdata -r [userwise_raw_data] -w [protobuf_binary] --method [protobuf] --size [int]

A sample of userwise raw data looks like:

0:
11,5.0
21,3.0
1:
9,5.0
12,1.0

where there are two users '0' and '1'.

Or if you only have rating wise raw data, you can first convert to a userwise raw data:

./getdata -r [rating_wise_raw] -w [userwise_raw_data] --method [userwise] --split [int]

A sample of rating_wise_raw data with a header looks like:

100000
0,1,5.0
0,2,1.0

where the header indicates the number of ratings, and follows by user_id, item_id and rating in each line.

Environment Requirment

• GCC 4.9 or higher

  tar zxf gcc-4.9.2.tar.gz;cd gcc-4.9.2;contrib/download_prerequisites;cd ..;mkdir buildc;cd buildc;../gcc-4.9.2/configure --disable-multilib;make -j 32;sudo make install;cd ..;

• Intel TBB

  sudo apt-get install libtbb-dev

• Google Protobuf

  sudo apt-get install -y libprotobuf-dev; sudo apt-get install -y protobuf-compiler;

• Intel MKL

Reference

[1] Fast Differentially Private Matrix Factorization. Ziqi Liu, Yu-Xiang Wang, Alex Smola.

[2] Learning recommender systems with adaptive regularization. Steffen Rendle.