MIND news recommadation using FM
Prerequisite
- python3.6+
- xlearn or other FM implementation
Clone this repo
git clone --recurse-submodules git@github.com:king0980692/mind.git
This Repo Structure
Usage
We collect the original dev set's impression results into truth.txt
, so we split this experiment into dev_part and test_part .
dev part
Using the result.txt
to evaulate the models result
test part
Using the official test data to evaulate the model result, you can upload your prediction to here
Script
đź’ˇ For both part , I write two script to run the experiment : dev_run.sh
and test_run.sh
Check them for detail !!