Recommendation technologies

Recommendation technologies are information technologies of providing information based on the collection, systematization and analysis of the data related to the preferences of Internet users located on the territory of the Russian Federation.

Rules for the use of recommendation technologies

  1. The database stores the history of user actions (voting) in the dating service system, including “like” and “dislike” marks. User filters installed in the dating service are also considered, namely: user location, gender, age, height, weight. All the above data is stored automatically in the dating service database.
  2. Accounting of user filters in the dating service occurs through the user interface, the operation of which is provided by server equipment that belongs to the dating service. The database, which stores the history of user actions (voting), also operates using server equipment of the dating service.
  3. Hidden factors characterizing user preferences are calculated on the dating service using matrix decomposition for previously sent votes (voting history) as part of user actions. The process of matrix decomposition and calculation of hidden factors occurs automatically. Based on the result of calculating hidden factors characterizing user preferences, recommendations on the most suitable profiles for the user are automatically generated. Consequently, the user receives the results of recommendations in the form of a list of profiles that corresponds to the user’s filters and, accordingly, his preferences.
  4. Matrix decomposition and calculation of hidden factors operate using program code.
  5. The data related to user preferences that is used to provide information using recommendation technologies includes: user actions in the “like” and “dislike” dating service system, as well as user filters: user location, gender, age, height, weight.
  6. The information related to user preferences can only be obtained by the dating service from the user.

JSC "Mamba"

All questions related to the use of recommendation technologies by Mamba JSC can be sent to: recomend@wamba.com