Skip to content

A recommendation system using matrix factorization algorithm with client application

Notifications You must be signed in to change notification settings

harteros/matrix-factorization-recommendation-system

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

4 Commits
 
 
 
 

Repository files navigation

matrix-factorization-recommendation-system

A recommendation system using matrix factorization algorithm with client application.

The currect application has 2 parts , the one is the backEnd and the other one is the frontEnd.

The backEnd directory inlcudes the matrix factorization algorithm which is broken down to multiple workers in order to complete the task faster.

In order to initiate the master - worker connection the user must specify the following things :

MasterNode class

port number to which the workers will connect
number of loops (seasons) for which the algorithm will run

WorkerNode class

port number at which the master has opened the connection
ip address of the masterNode (use localhost if running master and worker on the same pc)

Example files are given in order to test the algorithm (input_matrix_non_zeros.csv , POIs.json)

The frontEnd dictory contains the UI of the android application and it implements the training results from the MasterNode.
After the training is complete the user can ask for recommendations from the server through the ClientNode.

In order to do so the user must specify the port number to which the request will be sent and the ip adress of the server (masterNode).

Note 1 : localhost will not work here as the emulator of the android app is seen as an external device to the pc.
Note 2 : google maps api must be set in order to be able to view the maps (set api through file google_maps_api.xml)

About

A recommendation system using matrix factorization algorithm with client application

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages