Information recommendation method and system

Information recommendation method and system

  • CN 102,411,596 A
  • Filed: 09/21/2010
  • Published: 04/11/2012
  • Est. Priority Date: 09/21/2010
  • Status: Active Application
First Claim
Patent Images

1. an information recommendation method is characterized in that, comprising:

  • User behavior data in the inquiry preset time section;

    According to the user behavior data of being inquired about, obtain user'"'"'s subcategory interest-degree and/or user'"'"'s keyword interest-degree, and utilize subcategory interest-degree and/or the keyword interest-degree that is obtained, constitute user'"'"'s interest-degree vector;

    Utilize the vector angle formula, calculate active user'"'"'s interest-degree vector and the vectorial similarity of other user interest degree;

    According to similarity result of calculation, choose and at least one the highest user of active user'"'"'s interest-degree vector similarity, with selected user'"'"'s identification information, recommend to the active user;

    Wherein, said subcategory interest-degree is used to represent the relativeness of user to website subcategory visit interest-degree;

    Said keyword interest-degree is used to represent that the user uses the relativeness of interest to keyword.

View all claims

    Thank you for your feedback