drammos / MyLink

Web LinkedIn Application is exercise from our university, DIT - University of Athens.
MIT License
1 stars 0 forks source link

Add the Matrix Factorization algorithm for posts and jobs #51

Closed drammos closed 2 weeks ago

drammos commented 2 weeks ago

To do the bonus for jobs. Read the project

The user's timeline features articles based on a recommendation algorithm that will process a user's history of likes on articles other users, the submission of comments to them as well as the network of users. For her generation of the sentences used Matrix Factorization algorithm. So much for you this question as well as in question 7 technical data should be constructed as no free data is available. In case a user does not have noted interest and has not written any comments, the application will record and will leverages the projection of articles in order to construct the vectors that represent users in the article space. They will be shown on a priority basis articles whose authors belong to a user's network.

And for jobs: The appearance of the ads becomes so based on the skills it has declared by the user in the personal details of [Bonus] as well as based on the Matrix algorithm Factorization Collaborative Filtering that will process the network data of users. The vectors constructed represent the users in their space ads. Coordinate values ​​are determined by ad views from the users