A kind of in Web behavioral targeting, give user method and system for changeTechnical field
The present invention relates to web behavioral targeting, particularly when web behavioral targeting is for online advertisement, to losing markThe method again given for change of user of note, is specifically related to a kind of give user method and system for change in Web behavioral targeting.
Background technology
In web behavioral targeting is applied, visitor is marked by most crucial technology exactly, the most conventional labellingMethod is to use js cookie, and when user accesses website again, website can be by the information pair of storage in cookieUser demarcates and identifies.
But once user deletes cookie, or after hard disc of computer is reformatted, on subscriber computerCookie information will be lost, then when user accesses website again, and web behavioral targeting system just cannot be to userCarry out identification, can only be using this user as new user, then his historical behavior record and to his feature,The analysis results such as interest just cannot be mapped with him, say, that his historical data just fails.
Therefore, the method at present user being marked, once cookie can be made to lose, user is just difficult to be looked forReturning, all analysis results to the historical behavior of this user are the most all lost.This is one to the application of web behavioral targetingPlant the biggest loss.
Fig. 1 is the method giving user in web behavioral targeting for change of prior art: existing method is that website visiting is usedAccess website, family, website generates cookie labelling, contains unique identification ID of user in this cookie labellingNumber, in the computer of the implanted user of cookie tab file.Once the computer of user is formatted or cookieFile is deleted, and navigates to the link breakdown of website visiting user, and the most existing method cannot give user for change.
Summary of the invention
It is an object of the invention to, for overcoming in prior art once cookie to lose, user is just difficult to be retrieved,The analysis result of all historical behaviors to this user is the most all lost thus is caused the damage to the application of web behavioral targetingLose, thus provide a kind of multifactor in Web behavioral targeting to give user method for change.
For achieving the above object, the invention provides and a kind of in Web behavioral targeting, give user side for changeMethod, the method is for improving the discrimination power of web behavioral targeting system of users, it is characterised in that described methodComprise the steps of:
Step 101) make to generate factor when giving user for change by the habitual behavior pattern in user's history access behavior；
Step 102) by behavioural habits model value, the user losing cookie labelling is given for change again, thus shouldUser is mapped with the historic user being positioned, and has i.e. given the user losing cookie labelling for change so that webBehavioral targeting system can be repositioned onto this user.
In technique scheme, described habitual behavior pattern is:
According in a period of time n, user accesses the behavior number of times m of certain class website, Type of website t, acquired behavior mouldType h, then habitual behavior pattern h is:The habitual behavior pattern of user can determine: goes throughHistory h value h1, and current h2, ifIt is that h1 is corresponding that user corresponding for so h2 also gets over the most likely propertyUser.
In technique scheme, described behavior number of times m, this number of times m access, equal to user, the website that certain type is tNumber of times.
Present invention also offers based on said method and a kind of in Web behavioral targeting, give custom system for change, this systemFor to web behavioral targeting system is lost again the giving for change of user of cookie labelling, improving web behavioral targetingThe discrimination power of system of users, it is characterised in that described system comprises: access some users of website, online rowFor orientation subsystem and acquired behavior statistical model subsystem；
Described online behavioral targeting subsystem, does unique mark for the user to the website that maiden visit type is tID, this ID cookie identify file record, and the access behavior of record access user, described access behavior bagContain: the type of access is address and the time p of access of the website of t；
Described habitual behavior pattern statistics sub system, for exporting user to t type according to online behavioral targeting subsystemThe access time p of website, calculate interval of time n=p2-p1, access record further according to user accumulative,Draw access times m accessing t type website, the final habitual behavior pattern value calculating each access user.
In technique scheme, described habitual behavior pattern value is deposited in acquired behavior statistical model subsystem, withData base or document form persistence, and it is spaced p, regular update over time.
In technique scheme, described habitual behavior pattern statistics sub system comprises such as lower module further,
User habit behavior model computing module, this module accepts the input user every access record to t type websiteAccess time p, calculate interval of time n=p2-p1, access record according to user accumulative, draw accessAccess times m of t type website, the final habitual behavior pattern value calculating user；
The memory module of user habit behavior model value, user habit behavior model computing module is calculated by this moduleThe habitual behavior pattern value gone out carries out persistent storage；With
User's recovery module, this module, according to habitual behavior pattern value h, goes to retrieve h value in persistent storage identicalUser, and the ID of the user retrieved is exported in online behavioral targeting subsystem, by online behavioral targetingID is re-write in the cookie tab file of user by system.
It is an advantage of the current invention that proposition is a kind of when user loses or after deleting history cookie, again gives user for changeMethod.Once user loses or deleting history cookie, utilizes the method can realize again looking for userReturn, allow the historical record of this user and analysis result be mapped with this user, compensate for the most methodical deficiency,Enhance the effect of behavioral targeting.
Accompanying drawing explanation
Fig. 1 is the structural representation of the method giving user in web behavioral targeting for change of prior art；
Fig. 2 is the structural representation of the method giving user in web behavioral targeting for change of the present invention；
Fig. 3 is the flow chart giving user method in web behavioral targeting for change of the embodiment of the present invention.
Detailed description of the invention
With embodiment, the present invention is further described below in conjunction with the accompanying drawings.
Web behavioral targeting is by generating a cookie (tab file) on the user computer, remembers in cookieEmploy the ID at family, carry out unique certain user corresponding by ID.The most once User Format computer or user deleteExcept the cookie file on computer, then in web behavioral targeting, system just cannot position this user.Method describes: its method is by behavioural habits model algorithm, draws the behavioural habits model value h1 of certain new user a,If behavioural habits model value h1 is equal to or approximates value h2 of the behavioural habits model of certain historic user b, thenNew user a i.e. can corresponding historic user b, therefore, in web behavioral targeting, calculate even if depositing in user bCookie on machine is deleted, then user b still can be given for change by said method, and system is it is believed that a userIt is exactly the b user being deleted cookie.
Above-mentioned habitual behavior pattern: according in a period of time n, user accesses the behavior number of times m of certain class website,Type of website t, habitual behavior pattern h,The habitual behavior pattern of user may determine that: historyH value h1, and current h2, ifIt is use corresponding for h1 that user corresponding for so h2 also gets over the most likely propertyFamily.
As shown in Figure 2: website visiting user a, even if computer formats or cookie file is deleted, this methodStill habitual behavior pattern can be calculated according to the history access record of user a according to the algorithm of habitual behavior patternValue, it may be assumed that according in a period of time n, user accesses the behavior number of times m of certain class website, Type of website t, custom rowFor model h,Calculate habitual behavior pattern value h1 of user a.
Once there is website visiting user b, by calculating habitual behavior pattern value h2 of user b, if h2 is equal toH1 is then it is believed that user b i.e. loses the user a of cookie tab file, thus corresponding with user b by user aGet up, say, that system has given user a again for change.
Embodiment
As it is shown on figure 3, specifically comprise the following steps that
Step 101) computer to each website visiting user carries out the plantation of cookie tab file, cookieThe identity number ID, this ID that deposit labelling user in tab file uniquely identify a user.
Step 102) website records access user access behavior: Type of website t of access, the time p of access,According to time p, calculate interval of time n=p2-p1.Access record according to user accumulative, draw access tAccess times m of the Type of website, calculate habitual behavior pattern value
Step 103) once certain user a cookie labelling be deleted, i.e. label loss, then a period of time nAfter, according to h value h1 of user a, calculate the h value of all labeled users, as long as finding the custom of certain user bBehavior model value h2=h1, then it is believed that user b i.e. loses the user a of cookie tab file, thus will useFamily a and user b is mapped, say, that system has given user a again for change.
Step 104) ID in the cookie tab file of user b can be re-flagged the ID value into user a.So far, it is possible to successfully give user a for change.
It should be noted last that, above example is only in order to illustrate technical scheme and unrestricted.AlthoughWith reference to embodiment, the present invention is described in detail, it will be understood by those within the art that, to the present inventionTechnical scheme modify or equivalent, without departure from the spirit and scope of technical solution of the present invention, it is equalShould contain in the middle of scope of the presently claimed invention.