FIELD: computer technology.
SUBSTANCE: method for creating a personalized user parameter of interest implemented on a computer is disclosed. The method is implemented on a server, which is connected by means of a data transmission network to a client device connected to the user. The client device is made with the possibility of implementing a browser application, which previously had access through the data transmission network to at least one web-resource for a pre-determined previous time period. The method includes: obtaining by the server data on click history connected to the browser application, data on click history includes one or several uniform resource locators (hereinafter – URL) connected to corresponding one of at least one web-resource; creating based on data on click history one or several patterns of navigation clicks in a session, a pattern of navigation clicks in the session includes at least one of one or several URL, to which the browser application of the previous browser session had access; for each pattern of navigation clicks in the session, shortening each of one or several URL included in the pattern of navigation clicks in the session to obtain corresponding URL segment; creating by means of the first machine learning algorithm a corresponding vector value representing each of URL segments, wherein the first machine learning algorithm was educated to mark up URL segments for corresponding vector values based on a joint inclusion of URL segments in the pattern of navigation clicks in the session; assigning a weight value for each URL segment, wherein, for a given URL segment, the weight value is determined based on at least one of the first value, which is inversely proportional to a frequency of the given URL segment in URL segment log connected to URL, which were visited by server users, the second value, which is based on limitation period of access to URL connected to the given URL segment by the browser application; determining a value of user navigation profile based on at least one vector value and corresponding weight value, the value of navigation profile is connected to data on user click history for a pre-determined time period; creating by means of the second machine learning algorithm the personalized user parameter of interest connected to the user based on the value of navigation profile, wherein the second machine learning algorithm was educated to mark up the compliance of the value of user navigation profile with at least one personalized user parameter of interest.
EFFECT: technical result is creating the personalized user parameter of interest for providing delivery of a target content element for users.
21 cl, 7 dwg, 1 tbl
