Frequent Pattern Mining Engine
Engine helps by finding frequent patterns in training data. Usually used for complementary purchase cases but can be used to mine any patterns and propose complementary items.
Communication interface is based on PubNub queues.
How to use
docker pull goliasz/fpmaas-micro:v1.0
Once pulled just create a profile in PubNub. Find your PubKey and SubKey. Create your channel name and use this information for running the container.
docker run -dt --hostname fpm0 --name fpm0 -e "PN_PUBKEY=<YOUR PUB KEY>" -e "PN_SUBKEY=<YOUR SUB KEY>" -e "PN_CHANNEL=<YOUR CHANNEL NAME>" goliasz/fpmaas-micro:v1.0 /MyEngine/autostart.sh
Once container is started it is ready for training and serving subsequent items predictions.
Frequent Pattern Mining Engine implements similar PubNub communication protocol as Text Classification Engine. Training events has to be modeled in appriopriate way.