CN106126537A - Method and device is recommended in a kind of application - Google Patents

Abstract

The present invention provides a kind of application recommendation method, including: according to the internet records of each user, determine user to be recommended and the similar users of user to be recommended；Determine that similar users used but user to be recommended original application k, calculate similar users degree of adhesion and freshness to applying k；According to similar users to the application degree of adhesion of k, degree of association between freshness and similar users and user to be recommended, calculate the recommendation index of application k；And determine whether to recommend described application k for user to be recommended according to the recommendation index of described application k.Method is recommended in application provided by the present invention, it is possible to avoid by similar users recently not in use by application or use the lowest application of frequency to recommend user to be recommended, thus improve the availability being recommended to apply.

Description

An application recommendation method and device

technical field

The present invention relates to the field of information technology, in particular to an application recommendation method and device.

Background technique

With the rapid development of information technology, various application APPs (APPLITION mobile phone software applications) on mobile terminals have become the main channel for service providers to provide users with various value-added services. In the application stores of Apple and Google, the total amount of various applications Both exceeded one million, and consistently maintained a rapid growth trend. Faced with a large number of applications, users need to spend a lot of time to screen and try the applications before they can find the ones they are really interested in.

At present, telecom operators and service providers recommend applications of the same type based on the statistics of applications that users have used, and also based on the applications used by users who have the same interest in using them. However, this kind of recommendation to the user is based only on whether it is used or not, and the recommended application has great inaccuracy, for example, the usability of the recommended application is not high. For example, if you recommend an application that most users stop using soon after installation, in this case, the recommended application is inaccurate, the quality of the recommended application is not high, or the recommendation is invalid.

How to fundamentally solve the problem of accurately recommending applications for users, how to recommend highly available applications for users in combination with changes in users' interest in use, and improve the accuracy of application recommendation is an urgent problem in the field of information technology.

Contents of the invention

The technical problem to be solved by the present invention is to provide an application recommendation method and device for the above-mentioned defects in the prior art, to solve the problem in the prior art that it cannot be recommended according to the change of the user's interest. application problem.

To achieve the above purpose, the present invention provides an application recommendation method, including:

A method for recommending applications, the method comprising the steps of:

According to the online records of each user within the preset duration range, determine the user to be recommended and similar users, and the similar user is a user who uses the same application as the user to be recommended;

Determine the application k that has been used by similar users but has not been used by the user to be recommended, and calculate the adhesion and freshness of similar users to application k. The adhesion is the first time that similar users use the application within the preset time period The time difference between k and the last use of application k, the freshness is the time difference between the time when a similar user uses application k for the last time within the preset duration and the cut-off time of the preset duration;

Calculate the recommendation index of application k according to the adhesion and freshness of similar users to application k and the correlation between similar users and users to be recommended;

Whether to recommend the application k to the user to be recommended is determined according to the recommendation index of the application k and a preset first threshold.

Preferably, the determination of users to be recommended and similar users includes:

Divide the preset duration into at least two statistical periods, and calculate the traffic spent by each user using each application within each statistical period;

Calculate the time weight coefficient of each user corresponding to each statistical period;

Determine each user q who uses the same application j as the user to be recommended, and calculate the interest score of each user q for each application according to the traffic and time weight coefficient;

According to the interest score of the user to be recommended using application j, the average interest score of each application used by the user to be recommended, the interest score of user q using application j, and the average interest score of user q using each application, calculate the relationship between the user to be recommended and user q degree of relevance;

According to the degree of correlation between the user to be recommended and each user q, similar users to the user to be recommended are determined.

Preferably, in the calculation of the time weight coefficients of each user corresponding to each statistical period, the time weight coefficient f _ia (t) is calculated according to the following formula (1):

${\mathrm{<span\; class="notranslate">\; f</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; a</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; e</span>}}^{<span\; class="notranslate">\; \&lsqb;</span>\frac{{\mathrm{<span\; class="notranslate">\; t</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; a</span>}}}{{\mathrm{<span\; class="notranslate">\; t</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; A</span>}}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; \&rsqb;</span>}<span\; class="notranslate">\; ;</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>$

Among them, i means that the user to be recommended is user i;

t _iA is the number of time periods divided by user i;

ti _a is the ath time period of user i in the A time period;

That is, t _ia ∈ (1...t _ia ...t _iA ).

Preferably, in the calculation of the degree of adhesion and freshness of similar users to application k, the formula for calculating the degree of adhesion f _vk (t) is:

${\mathrm{<span\; class="notranslate">\; f</span>}}_{\mathrm{<span\; class="notranslate">\; vk</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; e</span>}}^{{\mathrm{<span\; class="notranslate">\; t</span>}}_{\mathrm{<span\; class="notranslate">\; vk</span>}\mathrm{<span\; class="notranslate">\; 1</span>}}}<span\; class="notranslate">\; ;</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; )</span>$

Among them, v is a similar user v;

t _vk1 is the time difference between the last time a similar user v visits application k and the time when he first visits application k.

Preferably, in the calculation of the affinity and freshness of similar users to application k, the calculation formula of freshness g _vk (t) is:

${\mathrm{<span\; class="notranslate">\; g</span>}}_{\mathrm{<span\; class="notranslate">\; vk</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; log</span>}}_{\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}}{\mathrm{<span\; class="notranslate">\; t</span>}}_{\mathrm{<span\; class="notranslate">\; vk</span>}\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; ;</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 3</span>}<span\; class="notranslate">\; )</span>$

Among them, v is a similar user v;

t _vk2 represents the minimum value of the time difference between the last visit to application k among similar users v and the current time.

The present invention also provides an application recommendation device, including:

The similar user module is used to determine the users to be recommended and similar users according to the online records of each user within a preset time range, and the similar users are users who use the same application as the users to be recommended;

Adhesion degree freshness module is used to determine the application k that has been used by similar users but has not been used by users to be recommended, and calculate the adhesion degree and freshness degree of similar users to application k. The time difference between the first use of application k and the last use of application k within the preset duration, the freshness is the time between the last time that similar users use application k within the preset duration and the preset duration cut-off time time difference;

The recommendation module is used to calculate the recommendation index of application k according to the degree of adhesion and freshness of similar users to application k, and the correlation between similar users and users to be recommended; Set the first threshold to determine whether to recommend the application k for the user to be recommended.

The similar user modules include:

A traffic calculation unit, configured to divide the preset duration into at least two statistical periods, and calculate the traffic spent by each user using each application within each statistical period;

The time weight unit is used to calculate the time weight coefficient of each user corresponding to each statistical period;

An interest scoring unit, configured to determine each user q who uses the same application j as the user to be recommended, and calculate the interest score of each user q for each application according to the traffic and time weight coefficient;

The correlation unit is used to calculate the to-be-recommended interest score based on the interest score of the user to be recommended using application j, the average interest score of each application used by the user to be recommended, the interest score of user q using application j, and the average interest score of user q using each application The degree of correlation between the user and user q, and according to the degree of correlation between the user to be recommended and each user q, determine similar users to the user to be recommended.

The time weight unit is specifically used to calculate the time weight coefficient, and the time weight coefficient f _ia (t) is calculated according to the following formula (1):

${\mathrm{<span\; class="notranslate">\; f</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; a</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; e</span>}}^{<span\; class="notranslate">\; \&lsqb;</span>\frac{{\mathrm{<span\; class="notranslate">\; t</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; a</span>}}}{{\mathrm{<span\; class="notranslate">\; t</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; A</span>}}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; \&rsqb;</span>}<span\; class="notranslate">\; ;</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>$

Among them, i means that the user to be recommended is user i;

t _iA is the number of time periods divided by user i;

t _ia is the ath time period of user i in the A time period;

That is, t _ia ∈ (1...t _ia ...t _iA ).

The viscosity freshness module is specifically used to calculate the viscosity f _vk (t), the viscosity

The formula for calculating the degree of fit f _vk (t) is:

${\mathrm{<span\; class="notranslate">\; f</span>}}_{\mathrm{<span\; class="notranslate">\; vk</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; e</span>}}^{{\mathrm{<span\; class="notranslate">\; t</span>}}_{\mathrm{<span\; class="notranslate">\; vk</span>}\mathrm{<span\; class="notranslate">\; 1</span>}}}<span\; class="notranslate">\; ;</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; )</span>$

Among them, v is a similar user v;

t _vk1 is the time difference between the last time a similar user v visits application k and the time when he first visits application k.

The adhesion freshness module is specifically used to calculate the freshness g _vk (t), and the calculation formula of the freshness g _vk (t) is:

${\mathrm{<span\; class="notranslate">\; g</span>}}_{\mathrm{<span\; class="notranslate">\; v</span>}\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; log</span>}}_{\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}}{\mathrm{<span\; class="notranslate">\; t</span>}}_{\mathrm{<span\; class="notranslate">\; v</span>}\mathrm{<span\; class="notranslate">\; k</span>}\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; ;</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 3</span>}<span\; class="notranslate">\; )</span>$

Among them, v is a similar user v;

t _vk2 represents the minimum value of the time difference between the last visit to application k among similar users v and the current time.

The application recommendation method and device provided by the present invention can recommend highly available applications for users to be recommended. Among the applications that the recommended user has never used, look for applications with high cohesiveness and high freshness of similar users to recommend to the users to be recommended. Degree means that the most recent time when similar users use the recommended application is closest to the recommended time. The application recommendation method provided by the present invention can avoid recommending applications that are no longer used by similar users or applications that are used very infrequently to users to be recommended, thereby improving the usability of the recommended applications.

Description of drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings that need to be used in the description of the embodiments. Obviously, the drawings in the following description are some embodiments of the present invention. For Those of ordinary skill in the art can also obtain other drawings based on these drawings without making creative efforts.

FIG. 1 is a schematic flowchart of the first embodiment of the application recommendation method provided by the present invention;

FIG. 2 is a schematic flowchart of a second embodiment of the application recommendation method provided by the present invention;

FIG. 3 is a schematic structural diagram of an application recommendation device provided by the present invention;

Fig. 4 is a schematic structural diagram of a similar user module in the application recommendation device provided by the present invention.

detailed description

In order to enable those skilled in the art to better understand the technical solutions of the present invention, the present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments. Apparently, the described embodiments are some, but not all, embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art without making creative efforts belong to the protection scope of the present invention.

Fig. 1 is a schematic flowchart of the first embodiment of the application recommendation method provided by the present invention. The application recommendation method shown in Fig. 1 includes the following steps:

Step S101, according to the online records of each user within a preset time range, determine the user to be recommended and similar users, the similar users are users who use the same application as the user to be recommended.

Specifically, only users who use the same application as the user to be recommended have the reference value of the recommended application. Among the applications used by the user to be recommended, at least one user with the same application can be considered as the user to be recommended similar users.

Step S102, determine the application k that has been used by similar users but not by the user to be recommended, and calculate the stickiness and freshness of similar users to application k, the stickiness being the first among similar users within the preset time period The time difference between the first use of application k and the last use of application k, the freshness is the time difference between the last time a similar user uses application k within the preset time period and the preset time length cut-off time.

Specifically, since it is necessary to exclude applications that the user has no longer used recently or that the frequency of use is too low, it is necessary to consider the adhesion and freshness of similar users for the application to be recommended. Wherein, the calculating formula of described degree of adhesion f _vk (t) is:

${\mathrm{<span\; class="notranslate">\; f</span>}}_{\mathrm{<span\; class="notranslate">\; vk</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; e</span>}}^{{\mathrm{<span\; class="notranslate">\; t</span>}}_{\mathrm{<span\; class="notranslate">\; vk</span>}\mathrm{<span\; class="notranslate">\; 1</span>}}}<span\; class="notranslate">\; ;</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; )</span>$

Among them, v is a similar user v;

t _vk1 is the time difference between the last time a similar user v visits application k and the time when he first visits application k.

The calculation formula of the freshness g _vk (t) is:

${\mathrm{<span\; class="notranslate">\; g</span>}}_{\mathrm{<span\; class="notranslate">\; v</span>}\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; log</span>}}_{\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}}{\mathrm{<span\; class="notranslate">\; t</span>}}_{\mathrm{<span\; class="notranslate">\; v</span>}\mathrm{<span\; class="notranslate">\; k</span>}\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; ;</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 3</span>}<span\; class="notranslate">\; )</span>$

Among them, v is a similar user v;

t _vk2 represents the minimum value of the time difference between the last visit to application k among similar users v and the current time.

After calculating the freshness and stickiness of user v for application k, statistics can be made on the usage of user v for application k.

Step S103, calculating the recommendation index of application k according to the degree of adhesion and freshness of similar users to application k and the correlation between similar users and users to be recommended.

Specifically, the calculation formula of the recommendation index provided by the present invention is:

${\mathrm{<span\; class="notranslate">\; P</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; =</span>\frac{{<span\; class="notranslate">\; \&Sigma;</span>}_{\mathrm{<span\; class="notranslate">\; v</span>}<span\; class="notranslate">\; \&Element;</span>\mathrm{<span\; class="notranslate">\; V</span>}\mathrm{<span\; class="notranslate">\; i</span>}}\mathrm{<span\; class="notranslate">\; the\; s</span>}\mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; m</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; v</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; v</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; -</span>\stackrel{<span\; class="notranslate">\; \&OverBar;</span>}{{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; v</span>}}}<span\; class="notranslate">\; )</span>{\mathrm{<span\; class="notranslate">\; g</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>{\mathrm{<span\; class="notranslate">\; f</span>}}_{\mathrm{<span\; class="notranslate">\; v</span>}\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>}{{<span\; class="notranslate">\; \&Sigma;</span>}_{\mathrm{<span\; class="notranslate">\; v</span>}<span\; class="notranslate">\; \&Element;</span>\mathrm{<span\; class="notranslate">\; V</span>}\mathrm{<span\; class="notranslate">\; i</span>}}\mathrm{<span\; class="notranslate">\; the\; s</span>}\mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; m</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; v</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; ;</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 6</span>}<span\; class="notranslate">\; )</span>$

Among them, i represents the user i to be recommended;

P _ik represents user i's recommendation index for application k;

V _i is a collection of similar users v of user i;

Indicates the average interest score of all applications used by user i;

Indicates the average interest score of user v using all its applications;

r _v,k represents the interest score of similar user v to application k;

sim(i,v) represents the correlation between user i and similar user v.

In order to make the calculation results of the recommendation index more intuitive and unified, the present invention also provides the following formula as another calculation formula of the recommendation index:

${{\mathrm{<span\; class="notranslate">\; P</span>}}^{<span\; class="notranslate">\; \&prime;</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; =</span>{\stackrel{<span\; class="notranslate">\; \&OverBar;</span>}{\mathrm{<span\; class="notranslate">\; r</span>}}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; +</span>\frac{{<span\; class="notranslate">\; \&Sigma;</span>}_{\mathrm{<span\; class="notranslate">\; v</span>}<span\; class="notranslate">\; \&Element;</span>\mathrm{<span\; class="notranslate">\; V</span>}\mathrm{<span\; class="notranslate">\; i</span>}}\mathrm{<span\; class="notranslate">\; the\; s</span>}\mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; m</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; v</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; v</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; -</span>\stackrel{<span\; class="notranslate">\; \&OverBar;</span>}{{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; v</span>}}}<span\; class="notranslate">\; )</span>{\mathrm{<span\; class="notranslate">\; g</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>{\mathrm{<span\; class="notranslate">\; f</span>}}_{\mathrm{<span\; class="notranslate">\; v</span>}\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>}{{<span\; class="notranslate">\; \&Sigma;</span>}_{\mathrm{<span\; class="notranslate">\; v</span>}<span\; class="notranslate">\; \&Element;</span>\mathrm{<span\; class="notranslate">\; V</span>}\mathrm{<span\; class="notranslate">\; i</span>}}\mathrm{<span\; class="notranslate">\; the\; s</span>}\mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; m</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; v</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; ;</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 7</span>}<span\; class="notranslate">\; )</span>$

Step S104, according to the recommendation index of the application k and a preset first threshold, determine whether to recommend the application k for the user to be recommended.

Specifically, according to the preset first threshold, the qualified application k is recommended to the user to be recommended. Alternatively, the recommendation index may be sorted, and an application k with a large recommendation index may be selected and recommended to the user to be recommended.

The application recommendation method provided by the present invention can recommend highly available applications for the user to be recommended. The specific implementation method is to find similar users who have the same application as the user to be recommended. Among the past applications, look for applications with high cohesion and high freshness of similar users to recommend to the users to be recommended. The last time you used the recommended app was closest to the recommended time. The application recommendation method provided by the present invention can avoid recommending applications that are no longer used by similar users or applications that are used very infrequently to users to be recommended, thereby improving the usability of the recommended applications.

Fig. 2 is a schematic flow chart of the second embodiment of the application recommendation method provided by the present invention. The application recommendation method shown in Fig. 2 includes the following steps:

Step S201, dividing the preset duration into at least two statistical periods, and calculating the traffic spent by each user using each application within each statistical period.

Specifically, the calculation of the traffic values of various applications used by each user includes all applications used by all users, where the traffic value is the traffic value of each application used by each user within the preset time range every day registration.

For example, the statistical period of a month can be divided into 6 statistical periods based on 5 days as a statistical period. However, in practice, since each user's use of the application is a subjective usage habit, the method of dividing the statistical period into a unified period may be This will cause deviations in the statistical results. Therefore, the method of dividing the statistical time period adopted in the present invention is to divide each user according to the time of the application used, so as to prevent the behavior of different users from being considered according to a unified standard.

To facilitate the subsequent detailed description of the solution, an example is given as follows. The default statistical period is one month, and the Internet access records of each user (only 3 users for example) using each application (only 2 applications for example) in January 2015 are extracted and calculated. The statistics of the traffic value and time of each user using various applications are shown in Table 1:

user name Application Name Flow value (M) time XXXX1 Youku 2 2015-1-1 XXXX1 Youku 5 2015-1-4 XXXX1 Youku 6 2015-1-6 XXXX1 Youku 3 2015-1-13 XXXX 2 Taobao 2 2015-1-1 XXXX 2 Taobao 6 2015-1-3 XXXX 2 Taobao 1 2015-1-12 XXXX 2 Youku 3 2015-1-20 XXXX 3 Taobao 6 2015-1-4 XXXX 3 Taobao 1 2015-1-6 XXXX 3 Taobao 3 2015-1-13

Table 1

In Table 1, according to the division standard of 5 days as a statistical period, user 1 can be divided into 3 statistical periods, user 2 can be divided into 4 statistical periods, and user 3 can be divided into 3 statistical periods.

Step S202, calculating the time weight coefficients of each user corresponding to each statistical period.

Specifically, in the calculation of the time weight coefficients of each user corresponding to each statistical period, the time weight coefficient f _ia (t) is calculated according to the following formula (1):

${\mathrm{<span\; class="notranslate">\; f</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; a</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; e</span>}}^{<span\; class="notranslate">\; \&lsqb;</span>\frac{{\mathrm{<span\; class="notranslate">\; t</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; a</span>}}}{{\mathrm{<span\; class="notranslate">\; t</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; A</span>}}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; \&rsqb;</span>}<span\; class="notranslate">\; ;</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>$

Among them, i means that the user to be recommended is user i;

t _iA is the number of time periods divided by user i;

t _ia is the ath time period of user i in the A time period;

That is, t _ia ∈ (1...t _ia ...t _iA ).

According to the calculation of formula (1),

The time weight coefficients of user 1 are 0.51, 0.72, 1;

The time weight coefficients of user 2 are 0.47, 0.60, 0.78, 1;

The time weight coefficients of user 3 are 0.51, 0.72, 1.

Step S203, determine each user q who uses the same application j as the user to be recommended, and calculate the interest score of each user q for each application according to the traffic and time weight coefficient.

Specifically, firstly, a user to be recommended is determined as user i, and users q who use the same application j as user i to be recommended can be used to calculate the degree of correlation with user i to be recommended. Otherwise, it is completely different from the application used by the user i to be recommended, and it is completely irrelevant to the user i to be recommended.

Specifically, the formula for calculating the interest score of user i using application j is:

${\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; =</span>\underset{\mathrm{<span\; class="notranslate">\; a</span>}<span\; class="notranslate">\; \&Element;</span>\mathrm{<span\; class="notranslate">\; A</span>}}{<span\; class="notranslate">\; \&Sigma;</span>}{\mathrm{<span\; class="notranslate">\; b</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; a</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; j</span>}}{\mathrm{<span\; class="notranslate">\; f</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; a</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; ;</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 4</span>}<span\; class="notranslate">\; )</span>$

Among them, b _ia is the traffic value of application j used by user i in the period a.

For example, the interest scores of users 1 to 3 using each application are calculated according to formula (4):

User 1's interest score for using Youku is: (2+5)*0.51+6*0.72+3*1=10.89;

User 2's interest score for using Taobao is: (2+6)*0.47+1*0.78＝4.54;

User 2's interest score for using Youku is: 3*1=3;

The interest score of user 3 in using Taobao is: 6*0.51+1*0.72+3*1=6.78.

If there are M total users in statistics, and N applications used by each user, then the interest score of each application used by each user can be represented by the following matrix:

$\left(\begin{array}{c}{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; 11</span>}}<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; 12</span>}}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; ...</span>}<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; ...</span>}<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}\mathrm{<span\; class="notranslate">\; N</span>}}\\ \mathrm{<span\; class="notranslate">\; ...</span>}\\ {\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; ...</span>}<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; ...</span>}<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; N</span>}}\\ \mathrm{<span\; class="notranslate">\; ...</span>}\\ {\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; m</span>}\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; m</span>}\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; ...</span>}<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; m</span>}\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; ...</span>}<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; m</span>}\mathrm{<span\; class="notranslate">\; N</span>}}\end{array}\right)$

Step S204, according to the interest score of the user to be recommended using application j, the average interest score of each application used by the user to be recommended, the interest score of user q using application j, and the average interest score of user q using each application, calculate the relationship between the user to be recommended and the user The correlation between q.

Specifically, the Pearson correlation coefficient formula is used to calculate the correlation between user i and user q, as shown in formula (5),

$\mathrm{<span\; class="notranslate">\; the\; s</span>}\mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; m</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; q</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\frac{{<span\; class="notranslate">\; \&Sigma;</span>}_{\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; \&Element;</span>\mathrm{<span\; class="notranslate">\; c</span>}\mathrm{<span\; class="notranslate">\; no</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; -</span>\stackrel{<span\; class="notranslate">\; \&OverBar;</span>}{{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; q</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; -</span>\stackrel{<span\; class="notranslate">\; \&OverBar;</span>}{{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; q</span>}}}<span\; class="notranslate">\; )</span>}{\sqrt{{<span\; class="notranslate">\; \&Sigma;</span>}_{\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; \&Element;</span>\mathrm{<span\; class="notranslate">\; c</span>}\mathrm{<span\; class="notranslate">\; no</span>}}{<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; -</span>\stackrel{<span\; class="notranslate">\; \&OverBar;</span>}{{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}}<span\; class="notranslate">\; )</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}\sqrt{{<span\; class="notranslate">\; \&Sigma;</span>}_{\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; \&Element;</span>\mathrm{<span\; class="notranslate">\; c</span>}\mathrm{<span\; class="notranslate">\; no</span>}}{<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; q</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; -</span>\stackrel{<span\; class="notranslate">\; \&OverBar;</span>}{{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; q</span>}}}<span\; class="notranslate">\; )</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}}<span\; class="notranslate">\; ;</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 5</span>}<span\; class="notranslate">\; )</span>$

Among them, q is each user q who uses the same application j as the user i to be recommended

Q is the set of users q, q∈Q;

Indicates the average interest score of all applications used by user i;

Indicates the average interest score of all applications used by user q;

cn represents the collection of applications j commonly used by user i and user q.

Step S205, according to the degree of correlation between the user to be recommended and each user q, determine similar users to the user to be recommended.

Specifically, after sorting the calculation results of the correlation degree in descending order, the user with the highest correlation degree can be selected as the similar user v of the user i to be recommended according to the preset screening rules.

It can be understood that similar user v is part or all of user q, and the above formula (5) is also applicable to the calculation of the similarity between user i and user v, namely:

$\mathrm{<span\; class="notranslate">\; the\; s</span>}\mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; m</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; v</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\frac{{<span\; class="notranslate">\; \&Sigma;</span>}_{\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; \&Element;</span>\mathrm{<span\; class="notranslate">\; c</span>}\mathrm{<span\; class="notranslate">\; no</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; -</span>\stackrel{<span\; class="notranslate">\; \&OverBar;</span>}{{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; v</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; -</span>\stackrel{<span\; class="notranslate">\; \&OverBar;</span>}{{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; v</span>}}}<span\; class="notranslate">\; )</span>}{\sqrt{{<span\; class="notranslate">\; \&Sigma;</span>}_{\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; \&Element;</span>\mathrm{<span\; class="notranslate">\; c</span>}\mathrm{<span\; class="notranslate">\; no</span>}}{<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; -</span>\stackrel{<span\; class="notranslate">\; \&OverBar;</span>}{{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}}<span\; class="notranslate">\; )</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}\sqrt{{<span\; class="notranslate">\; \&Sigma;</span>}_{\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; \&Element;</span>\mathrm{<span\; class="notranslate">\; c</span>}\mathrm{<span\; class="notranslate">\; no</span>}}{<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; v</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; -</span>\stackrel{<span\; class="notranslate">\; \&OverBar;</span>}{{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; v</span>}}}<span\; class="notranslate">\; )</span>}^{\mathrm{<span\; class="notranslate">\; 2</span>}}}}$

The application recommendation method provided by the present invention considers the change of the user's interest when determining similar users, and assigns different time weight coefficients to different time periods of the user. The earlier the time period, the smaller the weight coefficient, and the latest time period has the largest weight coefficient. , therefore, when a similar user no longer uses an application in the near future, the usage interest calculated based on the time weight and traffic value will be low, and the calculated correlation between the similar user and the user to be recommended will also be will be reduced accordingly, so that similar users who have the same interest changes as the user to be recommended can be found, and the usability of application recommendation can be improved.

Fig. 3 is a schematic structural diagram of an application recommendation device provided by the present invention, and the application recommendation device shown in Fig. 3 includes:

The similar user module 31 is configured to determine users to be recommended and similar users according to the online records of each user within a preset time range, and the similar users are users who use the same application as the user to be recommended.

Adhesion degree freshness module 32, is used for determining the application k that has been used by similar users but not used by users to be recommended, and calculates the degree of adhesion and freshness of similar users to application k. The time difference between the first use of application k and the last use of application k within the preset duration, and the freshness is the difference between the time when a similar user last used application k within the preset duration and the cut-off time of the preset duration time difference between

Specifically for calculating the degree of adhesion f _vk (t), the calculation formula of the degree of adhesion f _vk (t) is:

${\mathrm{<span\; class="notranslate">\; f</span>}}_{\mathrm{<span\; class="notranslate">\; vk</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; e</span>}}^{{\mathrm{<span\; class="notranslate">\; t</span>}}_{\mathrm{<span\; class="notranslate">\; vk</span>}\mathrm{<span\; class="notranslate">\; 1</span>}}}<span\; class="notranslate">\; ;</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; )</span>$

Among them, v is a similar user v;

t _vk1 is the time difference between the last time a similar user v visits application k and the time when he first visits application k.

Specifically for calculating the freshness g _vk (t), the calculation formula of the freshness g _vk (t) is:

${\mathrm{<span\; class="notranslate">\; g</span>}}_{\mathrm{<span\; class="notranslate">\; v</span>}\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; log</span>}}_{\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}}{\mathrm{<span\; class="notranslate">\; t</span>}}_{\mathrm{<span\; class="notranslate">\; v</span>}\mathrm{<span\; class="notranslate">\; k</span>}\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; ;</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 3</span>}<span\; class="notranslate">\; )</span>$

Among them, v is a similar user v;

t _vk2 represents the minimum value of the time difference between the last visit to application k among similar users v and the current time.

The recommendation module 33 is used to calculate the recommendation index of application k according to the degree of adhesion and freshness of similar users to application k, and the correlation between similar users and users to be recommended; and to calculate the recommendation index of application k according to the recommendation index and A preset first threshold is used to determine whether to recommend the application k for the user to be recommended.

The application recommendation device provided by the present invention can recommend highly available applications for users to be recommended. The specific implementation method is to find similar users who have the same interest in using applications as the users to be recommended, and find similar users who have used similar applications and users to be recommended Among the applications that have not been used, look for applications with high cohesion and freshness of similar users to recommend to users to be recommended. The cohesion is that similar users use the recommended application for a long time, and the freshness is The last time the similar users used the recommended application is closest to the recommended time. The application recommendation method provided by the present invention can avoid recommending applications that are no longer used by similar users or applications that are used very infrequently to users to be recommended, thereby improving the usability of the recommended applications.

FIG. 4 is a schematic structural diagram of a similar user module in the application recommendation device provided by the present invention. The similar user module in the application recommendation device shown in FIG. 4 includes:

The traffic calculation unit 311 is configured to divide the preset duration into at least two statistical periods, and calculate the traffic spent by each user using each application within each statistical period.

The time weight unit 312 is used to calculate the time weight coefficient of each user corresponding to each statistical period; specifically for calculating the time weight coefficient, the time weight coefficient f _ia (t) is calculated according to the following formula (1):

${\mathrm{<span\; class="notranslate">\; f</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; a</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; e</span>}}^{<span\; class="notranslate">\; \&lsqb;</span>\frac{{\mathrm{<span\; class="notranslate">\; t</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; a</span>}}}{{\mathrm{<span\; class="notranslate">\; t</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; A</span>}}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; \&rsqb;</span>}<span\; class="notranslate">\; ;</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>$

Among them, i means that the user to be recommended is user i;

t _iA is the number of time periods divided by user i;

t _ia is the ath time period of user i in the A time period;

That is, t _ia ∈ (1...t _ia ...t _iA ).

The interest score unit 313 is configured to determine each user q who uses the same application j as the user to be recommended, and calculate the interest score of each user q for each application according to the traffic and time weight coefficient.

The correlation unit 314 is configured to calculate the interest score of the user to be recommended using application j, the average interest score of each application used by the user to be recommended, the interest score of user q using application j, and the average interest score of user q using each application. Recommend the correlation between the user and the user q, and determine similar users of the user to be recommended according to the correlation between the user to be recommended and each user q.

The similar user module provided by the present invention considers the change of user interest when determining similar users, and assigns different time weight coefficients to different time periods of users. The earlier time period weight coefficients are smaller, and the most recent time period weight coefficients are the largest. , therefore, when a similar user no longer uses an application in the near future, the usage interest calculated based on the time weight and traffic value will be low, and the calculated correlation between the similar user and the user to be recommended will also be will be reduced accordingly, so that similar users who have the same interest changes as the user to be recommended can be found, and the usability of application recommendation can be improved.

In the several embodiments provided in this application, it should be understood that the disclosed method, device and system may be implemented in other ways. For example, the device embodiments described above are only illustrative, and the division of the functional modules is only a division of logical functions. In actual implementation, there may be other division methods, for example, multiple modules can be combined or can be Integrate into another system, or some features may be ignored, or not implemented.

Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: it can still be Modifications are made to the technical solutions described in the foregoing embodiments, or equivalent replacements are made to some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims (10)

1. An application recommendation method, characterized in that the method comprises the following steps: According to the online records of each user within the preset duration range, determine the user to be recommended and similar users, and the similar user is a user who uses the same application as the user to be recommended; Determine the application k that has been used by similar users but has not been used by the user to be recommended, and calculate the adhesion and freshness of similar users to application k. The adhesion is the first time that similar users use the application within the preset time period The time difference between k and the last use of application k, the freshness is the time difference between the time when a similar user uses application k for the last time within the preset duration and the cut-off time of the preset duration; Calculate the recommendation index of application k according to the adhesion and freshness of similar users to application k and the correlation between similar users and users to be recommended; Whether to recommend the application k to the user to be recommended is determined according to the recommendation index of the application k and a preset first threshold. 2. The application recommendation method according to claim 1, wherein said determining users to be recommended and similar users comprises: Divide the preset duration into at least two statistical periods, and calculate the traffic spent by each user using each application within each statistical period; Calculate the time weight coefficient of each user corresponding to each statistical period; Determine each user q who uses the same application j as the user to be recommended, and calculate the interest score of each user q for each application according to the traffic and time weight coefficient; According to the interest score of the user to be recommended using application j, the average interest score of each application used by the user to be recommended, the interest score of user q using application j, and the average interest score of user q using each application, calculate the relationship between the user to be recommended and user q degree of relevance; According to the degree of correlation between the user to be recommended and each user q, similar users to the user to be recommended are determined. 3. The application recommendation method according to claim 2, wherein, in the calculation of the time weight coefficients corresponding to each statistical period for each user, the time weight coefficient f _ia (t) is calculated according to the following formula (1): ${\mathrm{<span\; class="notranslate">\; f</span>}}_{{\mathrm{<span\; class="notranslate">\; i</span>}}_{\mathrm{<span\; class="notranslate">\; a</span>}}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; e</span>}}^{<span\; class="notranslate">\; \&lsqb;</span>\frac{{\mathrm{<span\; class="notranslate">\; t</span>}}_{{\mathrm{<span\; class="notranslate">\; i</span>}}_{\mathrm{<span\; class="notranslate">\; a</span>}}}}{{\mathrm{<span\; class="notranslate">\; t</span>}}_{{\mathrm{<span\; class="notranslate">\; i</span>}}_{\mathrm{<span\; class="notranslate">\; A</span>}}}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; \&rsqb;</span>}<span\; class="notranslate">\; ;</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>$ Among them, i means that the user to be recommended is user i; t _iA is the number of time periods divided by user i; t _ia is the ath time period of user i in the A time period; That is, t _ia ∈ (1...t _ia ...t _iA ). 4. The application recommendation method as claimed in claim 1, wherein, in the calculation of the degree of adhesion and freshness of similar users to application k, the formula for calculating the degree of adhesion f _vk (t) is: f _vk (t) = e ^tvk1 ; (2) Among them, v is a similar user v; t _vk1 is the time difference between the last time a similar user v visits application k and the time when he first visits application k. 5. The application recommendation method according to claim 1, wherein, in the calculation of the degree of adhesion and freshness of similar users to application k, the calculation formula of freshness g _vk (t) is: ${\mathrm{<span\; class="notranslate">\; g</span>}}_{\mathrm{<span\; class="notranslate">\; v</span>}\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; log</span>}}_{\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}}{\mathrm{<span\; class="notranslate">\; t</span>}}_{\mathrm{<span\; class="notranslate">\; v</span>}\mathrm{<span\; class="notranslate">\; k</span>}\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; ;</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 3</span>}<span\; class="notranslate">\; )</span>$ Among them, v is a similar user v; t _vk2 represents the minimum value of the time difference between the last visit to application k among similar users v and the current time. 6. An application recommendation device, characterized in that it comprises: The similar user module is used to determine the users to be recommended and similar users according to the online records of each user within a preset time range, and the similar users are users who use the same application as the users to be recommended; Adhesion degree freshness module is used to determine the application k that has been used by similar users but has not been used by users to be recommended, and calculates the degree of adhesion and freshness of similar users to application k, the degree of adhesion is the similar user in the The time difference between the first use of application k and the last use of application k within the preset time period, the freshness is the time between the last time a similar user uses application k within the preset time period and the preset time length cut-off time time difference; The recommendation module is used to calculate the recommendation index of application k according to the degree of adhesion and freshness of similar users to application k, and the correlation between similar users and users to be recommended; Set the first threshold to determine whether to recommend the application k for the user to be recommended. 7. The application recommendation device according to claim 6, characterized in that: The similar user modules include: A traffic calculation unit, configured to divide the preset duration into at least two statistical periods, and calculate the traffic spent by each user using each application within each statistical period; The time weight unit is used to calculate the time weight coefficient of each user corresponding to each statistical period; An interest scoring unit, configured to determine each user q who uses the same application j as the user to be recommended, and calculate the interest score of each user q for each application according to the traffic and time weight coefficient; The correlation unit is used to calculate the to-be-recommended interest score based on the interest score of the user to be recommended using application j, the average interest score of each application used by the user to be recommended, the interest score of user q using application j, and the average interest score of user q using each application The degree of correlation between the user and user q, and according to the degree of correlation between the user to be recommended and each user q, determine similar users of the user to be recommended. 8. The application recommendation device according to claim 7, characterized in that: The time weight unit is specifically used to calculate the time weight coefficient, and the time weight coefficient f _ia (t) is calculated according to the following formula (1): ${\mathrm{<span\; class="notranslate">\; f</span>}}_{{\mathrm{<span\; class="notranslate">\; i</span>}}_{\mathrm{<span\; class="notranslate">\; a</span>}}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; e</span>}}^{<span\; class="notranslate">\; \&lsqb;</span>\frac{{\mathrm{<span\; class="notranslate">\; t</span>}}_{{\mathrm{<span\; class="notranslate">\; i</span>}}_{\mathrm{<span\; class="notranslate">\; a</span>}}}}{{\mathrm{<span\; class="notranslate">\; t</span>}}_{{\mathrm{<span\; class="notranslate">\; i</span>}}_{\mathrm{<span\; class="notranslate">\; A</span>}}}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; \&rsqb;</span>}<span\; class="notranslate">\; ;</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>$ Among them, i means that the user to be recommended is user i; t _iA is the number of time periods divided by user i; t _ia is the ath time period of user i in the A time period; That is, t _ia ∈ (1...t _ia ...t _iA ). 9. The application recommendation device according to claim 6, characterized in that: The degree of adhesion freshness module is specifically used to calculate the degree of adhesion f _vk (t), and the calculation formula of the degree of adhesion f _vk (t) is: f _vk (t) = e ^tvk1 ; (2) Among them, v is a similar user v; t _vk1 is the time difference between the last time a similar user v visits application k and the time when he first visits application k. 10. The application recommendation device according to claim 6, characterized in that: The adhesion freshness module is specifically used to calculate the freshness g _vk (t), and the calculation formula of the freshness g _vk (t) is: ${\mathrm{<span\; class="notranslate">\; g</span>}}_{\mathrm{<span\; class="notranslate">\; v</span>}\mathrm{<span\; class="notranslate">\; k</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; log</span>}}_{\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; 2</span>}}}{\mathrm{<span\; class="notranslate">\; t</span>}}_{\mathrm{<span\; class="notranslate">\; v</span>}\mathrm{<span\; class="notranslate">\; k</span>}\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; ;</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 3</span>}<span\; class="notranslate">\; )</span>$ Among them, v is a similar user v; t _vk2 represents the minimum value of the time difference between the last visit to application k among similar users v and the current time.