The ffm_mpi from jiarunying

ffm_mpi's Introduction

Introduction

field-aware factorization machie[1] solve by ftrl algorighm[2], based on mpich
Dependency

openblas

mpich-2

Use

This folder shoulder in same path on all of your cluseter nodes.

all of follow action in the folder of Field-aware-Factorization-Machine-ftrl-mpi step 1:

 split your data into n parts. (n equal the number of nodes of your cluster), 
 for example, cd ./data and run the command: sh run_split_data.sh

step 2:

 cd ..

 modeify Ip in the file run_ffm_mpi.sh and then rum commend: sh run_ffm_mpi.sh

Note:

 if you want to stop the programme, run commend: sh stop.sh

Feature list

4.1. About model selection:

 4.1.1. if you want to use LR(Logistic Regression) model only, set the parameter in config.h: isffm=0 isfm=0 islr=1

 4.1.2. if you want to use FM(Factorization Machine) model only, set the parameter in config.h: isffm=0 isfm=1 islr=0

 4.1.3. if you want to use FFM(Field-aware Factorization Machine) model only, set the parameter in config.h: isffm=1 isfm=0 islr=0

4.2. Evaluation

 print AUC after some epochs

4.3 Save Model

dump model to the model folder

Todo list

1, multithread

2, same model paralle on Parameter Server

3, load data batch

4, optimazation the learning algorithm
Contact:
```
 [email protected]
```
References:

[1] Field-aware Factorization Machines for CTR Prediction. http://www.csie.ntu.edu.tw/~cjlin/papers/ffm.pdf

[2] Ad Click Prediction: a View from the Trenches. http://dl.acm.org/citation.cfm?id=2488200

[3] Factorization Machine. http://www.csie.ntu.edu.tw/~b97053/paper/Rendle2010FM.pdf