• InFortis News

  • картинка – InFortisNews

  • InFortis News

    We provide personalized recommendations of the content for users visiting the publisher websites and informational and entertaining platforms.

    What’s new?

    Our approach differs from the others by analysis of a user’s personality. A user gets the content that corresponds to his actual needs, his outlook, his unique preferences and information consumption traits.

    Recommendation performance

    According to www.tut.by our service provided 26% CTR increase comparing to other recommendation algorithms used by the website (context, popular news). For reference – context approach in recommendation shows 1,9% CTR increase only comparing to occasional recommendation of the content.

    Such performance results are achieved by means of our algorithms based on human mental modelling technology.

    Studying consistent patterns of human mind performance we have an opportunity to correlate outer behavior of a man (as for the Internet it includes clicks on article and news links) with some processes in his inner world. It permits to form a mental model for every website user.

    As soon as accumulated information about a user’s mentality is enough to model his behavior the system launches the process. Behavior we can predict in this way includes what content a user chooses.

    What’s required?

    We find a mental model based on user’s 20 activity pieces (clicks) well enough for further operation. Upon overcoming this limit a user starts to receive our recommendations. As far as our system is dynamic, – it updates information about a user after he has clicked a link, – then our recommendations become more and more accurate with every visit of the given user.


    It takes 0,1 sec for the system to form a recommendation list. It is realized while a web page is being loaded. All calculations are carried out by our side.

    A website owner can provide a user for recommendations in the separate news box but we can also personalize the news feed or the news inside topic sections.

    It does not matter how many users you have – we do not use collaborative filtering, we handle each user personally.

    Other benefits

    Analysis of readers’ activity provides a base we can use to solve other tasks. It means we can collaborate in other fields (plz, see other InFortis projects).

    Advertising prospects are important for website owners. Using mental models based on website data we can target users by needs and goods, personalize advertisements, etc. For further information about our service in this field, please, visit InFortis Marketing.