Commodity recommendation method and device

Commodity recommendation method and device

  • CN 105,653,657 A
  • Filed: 12/25/2015
  • Published: 06/08/2016
  • Est. Priority Date: 12/25/2015
  • Status: Active Application
First Claim
Patent Images

1. the recommend method of commodity, it is characterised in that, described recommend method comprises:

  • Obtaining rating matrix, described rating matrix comprises multiple user to the actual scoring of multiple commodity, and fills up randomized number in the vacant locations of described rating matrix;

    For each user, the implicit feedback data corresponding according to described rating matrix and described user, carrying out successive ignition by SVD++ algorithm to solve, obtain the scoring vector that described user is corresponding, described scoring vector comprises described user and the prediction of each commodity is marked;

    When recommending commodity to user, the scoring vector corresponding according to described user, chooses the maximum commercial product recommending of prediction scoring to described user.

View all claims
    ×
    ×

    Thank you for your feedback

    ×
    ×