Changes between Version 1 and Version 2 of ClassificationOfUsersForAds


Ignore:
Timestamp:
Jul 14, 2017, 10:58:38 AM (3 years ago)
Author:
psantos
Comment:

--

Legend:

Unmodified
Added
Removed
Modified
  • ClassificationOfUsersForAds

    v1 v2  
    44
    55== Project Details ==
    6 
     6On this report, a proposal for the novel content targeting system is
     7presented. The proposed system can be a general user classification
     8mechanism that would allow us to scale app recommendations for
     9groups of users with similar interests. This system could be applied to
     10the organic recommendations by the Editorial and Community teams,
     11but also to the Advertising/Monetisation processes.
     12The system is partially automated, since it allows us to understand each
     13user group qualitatively and manually manage the recommendations for
     14each group over time. Part of these recommendations can be
     15automated based on simple rules such as app categories, trending apps
     16within each group and most downloaded but not yet installed apps per
     17group. It is considered that this combined marketing research/data
     18science approach is the best solution to implement a recommender
     19system in a very short term. The solution is also scalable and can be
     20tweaked and improved over time.
     21The proposed system can be constructed based on a simple three stage
     22data mining process. The steps are described in detail from a
     23mathematical point of view, and some technological implications are
     24also addressed. Therefore this report can serve as the roadmap for the
     25implementation of the system.
    726=== Purpose of this project ===
    827