CN105930507A - Method and apparatus for obtaining Web browsing interest of user - Google Patents

Abstract

The invention discloses a method for obtaining a Web browsing interest of a user. The method is applied to a data processing device installed with a program with a webpage browsing function. The method comprises the steps of judging whether a condition of computing the browsing interest is met or not, and if yes, reading a user browsing log between a moment of previously computing the browsing interest and a current moment; obtaining an interest tag of a webpage corresponding to the read user browsing log and a type that the interest tag belongs to; computing a weight of the interest tag of each type; and for the interest tag of any type, multiplying a interest computing value of the previously computed interest tag of the type by an attenuation factor, and accumulating product and the currently computed weight of the interest tag of the type to obtain an interest computing value of the currently computed interest tag of the type, storing the interest computing value of the currently computed interest tag of the type, and returning to the step of judging whether the condition of computing the browsing interest is met or not. The invention furthermore discloses an apparatus for obtaining the Web browsing interest of the user.

Description

Method and device for obtaining Web browsing interest of user

Technical Field

The present application relates to the field of internet technologies, and in particular, to a method and an apparatus for obtaining a Web browsing interest of a user.

Background

In recent years, with the rapid development of the internet, the Web information is rapidly increased in a geometric progression, and people often need to spend a lot of time to find the required information on the internet. In this case, personalized recommendation is an effective method for solving the current information overload problem as an important means for information filtering. And the user interest model is the core of the personalized recommendation system. A good interest model can better improve the internet experience and the information utilization efficiency of users, so the research of the interest model becomes a hot topic of academia and IT.

Because the data of the user historical behaviors is too large, the browsing interest calculation method commonly adopted in the prior art comprises the following steps: and when the user browsing interest is calculated every day, only the user browsing behavior of one time window can be selected from the historical log for calculation. This way of calculation is not only computationally intensive, but also loses the influence of user interest outside the selected time window.

Disclosure of Invention

The application provides a method and a device for obtaining the Web browsing interest of a user, which can obtain the Web browsing interest of the user with a small calculation amount.

An embodiment of the present application provides a method for obtaining a Web browsing interest of a user, the method being applied to a data processing apparatus in which a program having a Web browsing function is installed, and the method including:

judging whether the condition of calculating the browsing interest is met, if so, reading a user browsing log from the last time of calculating the browsing interest to the current time;

obtaining interest tags of web pages corresponding to the read user browsing logs and categories to which the interest tags belong;

respectively calculating the weight of the interest tag of each category;

for any category of interest tags, multiplying the last calculated interest calculation value of the category of interest tags by the attenuation factor related to time, then accumulating the obtained value with the weight of the currently calculated interest tags of the category to obtain the interest calculation value of the currently calculated interest tags of the category, storing the currently calculated interest calculation value of the interest tags of the category, and returning to the step of judging whether the condition of calculating browsing interest is met.

Another embodiment of the present application also provides an apparatus for obtaining a Web browsing interest of a user, the apparatus being located on a data processing device on which a program having a Web browsing function is installed, the apparatus including: the system comprises a condition judgment module, a reading module, a weight calculation module, an interest calculation module and a storage module;

the condition judgment module is used for judging whether the condition for calculating the browsing interest is met, and if so, the reading module, the weight calculation module and the interest calculation module are enabled;

the reading module is used for reading a user browsing log between the last time of calculating browsing interest and the current time, and acquiring an interest tag of a webpage corresponding to the read user browsing log and a category to which the interest tag belongs;

the weight calculation module is used for calculating the weight of each category of interest tag acquired by the reading module respectively;

the interest calculation module is used for multiplying the interest calculation value of the interest label of the category, which is stored in the storage module at the last time, of the interest label of the category by a time-dependent attenuation factor, then accumulating the interest calculation value of the interest label of the category, which is calculated at the present time, of the weight calculation module to obtain the interest calculation value of the interest label of the category, which is calculated at the present time, and sending the interest calculation value of the interest label of the category, which is calculated at the present time, to the storage module for storage;

the storage module is used for storing the interest calculation value from the interest calculation module.

According to the technical scheme, the calculation result of the Web browsing interest is only related to the previous calculation result and the browsing behavior of the user in the time period from the previous time to the current time, so that the Web browsing interest of the user can be obtained by adopting an incremental calculation mode, the calculation amount is greatly reduced, the historical behavior of the user can be accumulated, and the influence of time factors is comprehensively considered.

Drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings needed to be used in the description of the embodiments are briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without creative efforts. Wherein,

FIG. 1 is a schematic diagram of an implementation environment in accordance with an embodiment of the present invention;

fig. 2 is a flowchart illustrating a method for obtaining a Web browsing interest of a user according to an embodiment of the present invention;

FIG. 3 is a schematic diagram of an apparatus for obtaining a Web browsing interest of a user according to an embodiment of the present invention;

fig. 4 is a schematic structural diagram of a data processing device according to an embodiment of the present invention.

Detailed Description

The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

Fig. 1 is a schematic diagram of an implementation environment according to an embodiment of the invention. Referring to fig. 1, a user terminal 103 is connected to at least one Web server 101 through the internet 102 and browses page contents provided by the Web server 101. Wherein different pages carry type tags associated with the content of the page. For example, a live page of an NBA ball game, the type tag of which may include: basketball, sports, NBA, etc.

The user terminal 103 generates a browsing log according to the history of the user browsing the web pages, and records the related information of the web pages browsed by the user, wherein the related information at least comprises browsing time and type tags of the web pages.

The user terminal 103 may be a data processing apparatus in which a web browser program or a program having a web browsing function (e.g., a news client) is installed. The data processing device may be a mobile terminal or a PC, the mobile terminal including but not limited to a smartphone, a tablet computer, etc. The PC may be a personal computer, a notebook, a super book, etc.

The scheme of the invention adopts an incremental calculation method to obtain the Web browsing interest of the user, the Web browsing interest is calculated every preset time, and the calculation result is only related to the last calculation result and the browsing behavior of the user in the time period from the last time to the current time.

Embodiments of the present invention provide a method for obtaining a user's Web browsing interest, assuming that the weight of the user's browsing interest is only related to the number of times (i.e., PV) he (she) browses a certain type of Web page, and the influence of the PV of that type of interest on the interest weight is attenuated in time. According to this assumption, when calculating the browsing interest weight of the user, the influence of PV in the recent period of time is more focused, and the long-term interest of the user is focused. That is, when a user is considered to be interested in a certain type of web page, the user keeps paying attention to the type of web page, and the user frequently browses the type of web page.

The embodiment of the invention provides a method flow for obtaining the Web browsing interest of a user, which is executed by a user terminal and comprises the following steps as shown in figure 2:

step 201: and judging whether the condition for calculating the browsing interest is met, if so, continuing to execute the step 202.

The condition for calculating the browsing interest may be various, for example, if the set rule is that the browsing interest is calculated every predetermined time, the condition is that the current system time is a preset time. The preset time may be set to a time period with a low probability of occurrence of the browsing behavior, for example, set between 23 a and 5 a in the morning according to the usage habits of general users. The interval between adjacent preset time instants may be 1 day, or may be other time intervals longer or shorter, such as 2 days, half days, etc.

If the user terminal is powered on, the preset time is after the power is off and before the current time, the condition for calculating the browsing interest is also considered to be met.

Step 202: and reading a user browsing log from the last time of calculating the browsing interest to the current time.

Step 203: and acquiring the interest tags of the web pages corresponding to the read user browsing logs and the categories of the interest tags.

Step 204: and respectively calculating the weight of the interest label of each category.

The interest labels can be divided into different categories, such as finance categories, sports categories, fashion entertainment categories, fitness and health categories, and weights of the interest labels are calculated for the different categories respectively, so that browsing interests of users for the different categories can be known. The weight is related to the web browsing with the interest tag of the type from the last time the browsing interest is calculated to the current time. Specific calculation examples will be given later.

Step 205: for any category of interest tags, the last calculated interest calculation value of the category of interest tags is multiplied by the attenuation factor related to time and then is accumulated with the weight of the currently calculated interest tags of the category to obtain the interest calculation value of the currently calculated interest tags of the category, the currently calculated interest calculation value of the category of interest tags is stored, and the step 201 is returned.

If the browsing interest is currently calculated for the first time, the interest calculation value of the interest tag calculated last in this step is zero.

For the implementation mode of calculating the browsing interest once a day, one embodiment of the invention provides a specific interest value expression:

$w\left(n\right)=\underset{i=0}{\overset{n-1}{\Σ}}k*\ln ({\mathrm{PV}}_i+1)\·\exp (-\mathrm{\λ}i)---\left(1\right)$

wherein:

w (n) is an interest calculation value of a certain type of label within past n days until the current calculation time.

PV_iThe number of browsing times of the web pages of the interest tags on the ith day from the current calculation time is shown.

k is a normalized coefficient of the interest weight. Since the average PV of the user is different for different interests, different k values can be determined according to different interests to adjust ln (PV)_iThe degree of influence of +1) is,

λ is a time-dependent attenuation factor that can be derived from a formula of half-life (the value of interest is reduced to half of the initial value). The half-life is given by:

$t_{1/2}=\frac{\ln 2}{\mathrm{\λ}}---\left(2\right)$

if the half-life is 60 days, λ is 0.012. It is clear that if the half-life takes other values, the time dependent attenuation factor changes accordingly.

Let the interest value at t0 be:

$w\left(n\right)=\underset{i=0}{\overset{n-1}{\Σ}}\ln ({\mathrm{PV}}_i+1)\·\exp (-\mathrm{\λ}i)$

for the same interest, k is the same, and for the simplicity of the derivation process, k is not considered in the derivation below.

Then the interest value for t0+1 day is:

$\begin{array}{c}{w}^{,}\left(n\right)=\underset{i=0}{\overset{n}{\Σ}}\ln ({\mathrm{PV}}_i+1)\·\exp (-\mathrm{\λ}i)\\ =\[\underset{i=0}{\overset{n-1}{\Σ}}\ln ({\mathrm{PV}}_i+1)\·\exp (-\mathrm{\λ}i)\]\·\exp (-\mathrm{\λ})+\ln ({\mathrm{PV}}_0+1)\\ =w\left(n\right)\·\exp (-\mathrm{\λ})+\ln ({\mathrm{PV}}_0+1)\end{array}---\left(3\right)$

the first term to the right of the middle sign in the formula (3) is the interest calculation value at t0 days multiplied by the attenuation factor, and the second term is the interest label weight at t0+1 days, which shows that the interest label weight of today only relates to the calculation result at yesterday and the interest label PV of today, so that the incremental calculation of the browsing interest can be performed.

According to this calculation, the long-term interest is much larger than the short-term explosive interest. Taking lambda as 0.012, such as:

within 10 days, if the click rate of the user a to the interest a is 10 every day, the interest value is: 21.82;

if the user B has not clicked on the interest a in the previous 9 days, and the click amount is 100 today, the interest value is as follows: 4.08.

although the total click volume for interest a is consistent between user A and user B within 10 days, the calculated value for interest a is much larger than that for user B, which is also more practical.

For a real user, there is a limit to the interest value calculated by the calculation formula. Assuming that the user a is very enthusiastic about the interest a, the browsing amount per day is 1000 times (the actual probability is very small), and after the infinite days, the interest value is:

$w=\underset{n->\mathrm{\∞}}{\lim }\underset{i=0}{\overset{n-1}{\Σ}}\ln (1000+1)\·\exp (-\mathrm{\λ})=\underset{n->\mathrm{\∞}}{\lim }\frac{\ln 1001(1-\exp (-\mathrm{\λ}n))}{1-\exp (-\mathrm{\λ})}$

if λ is 0.012, then: and w is 579.11.

From the formula and derivation designed above, it can be seen that the browsing interest calculation value of a certain category of a user on a certain day is only related to the browsing behavior of the interest tag webpage of the category of the user on the same day and the browsing interest calculation value of the category of the user on yesterday, an incremental calculation mode can be adopted, and the influence of the historical browsing behavior of the user is included, and meanwhile, the influence of the behavior which is attenuated according to time is also considered.

An embodiment of the present invention further provides an apparatus for obtaining a Web browsing interest of a user, where the apparatus is located in a data processing device installed with a Web browser program or a program with a Web browsing function, as shown in fig. 3, the apparatus 300 includes: a condition judgment module 301, a reading module 302, a weight calculation module 303, an interest calculation module 304 and a storage module 305;

the condition judgment module 301 is used for judging whether the condition for calculating the browsing interest is met, and if the condition is met, the reading module 302, the weight calculation module 303 and the interest calculation module 304 are enabled;

the reading module 302 is configured to read a user browsing log from a previous time when the browsing interest is calculated to a current time, and acquire an interest tag of a webpage corresponding to the read user browsing log and a category to which the interest tag belongs;

a weight calculating module 303, configured to calculate a weight of each category of interest tag acquired by the reading module;

an interest calculation module 304, configured to, for any type of interest tag, multiply the attenuation factor by the last calculated interest calculation value of the type of interest tag stored in the storage module 305, and then accumulate the result with the weight of the type of interest tag calculated by the weight calculation module this time to obtain the interest calculation value of the type of interest tag calculated this time, and send the calculated interest calculation value of the type of interest tag calculated this time to the storage module 305 for storage;

the storage module 305 is used for storing the interest calculation value from the interest calculation module.

According to another embodiment of the present invention, the condition for calculating the browsing interest is: the current system time is a preset time.

According to another embodiment of the present invention, there are a plurality of preset times, and the interval between adjacent preset times is at least one day; the conditions for calculating the browsing interest are as follows: the current system time is one of the preset moments. The interval between the adjacent preset moments is one day, the browsing interest is calculated for the (n +1) th time, and the expression of the calculated interest value obtained by the interest calculating module is as follows:

w'(n)＝w(n)·exp(-λ)+ln(PV₀+1)

wherein w' (n) is the calculated interest value on day n +1, and w (n) is the calculated interest value on day n; PV (photovoltaic)₀The number of views of the web page of the category interest tag on day n +1, ln (PV)₀+1) is the weight of the interest tag of the category calculated on day n + 1; λ is a time dependent attenuation factor.

According to another embodiment of the present invention, if the data processing apparatus is powered on, the preset time is after the data processing apparatus is powered off and before the current time, and the condition determining module 301 determines that the condition for calculating the browsing interest is satisfied.

In addition, each module in the embodiments of the present invention may be integrated into one processing unit, or each module may exist alone physically, or two or more modules may be integrated into one unit. The integrated unit can be realized in a form of hardware, and can also be realized in a form of a software functional unit.

In an embodiment, the data processing apparatus 400 may include: memory 410, processor 411, bus 412, port 413. The processor 411 and the memory 410 are interconnected by a bus 412. Processor 411 may receive and transmit data through port 413 to enable network communications. Each of the modules 301-305 may be machine-executable instructions stored in the memory 410. The processor 411 can implement the functions of the modules 301 to 305 by executing machine-executable instructions contained in the modules 301 to 305 in the memory 410.

In addition, each embodiment of the present invention can be realized by a data processing program executed by a data processing apparatus such as a computer. It is clear that the data processing program constitutes the invention. Further, the data processing program, which is generally stored in one storage medium, is executed by directly reading the program out of the storage medium or by installing or copying the program into a storage device (such as a hard disk and/or a memory) of the data processing device. Such a storage medium therefore also constitutes the present invention. The storage medium may use any type of recording means, such as a paper storage medium (e.g., paper tape, etc.), a magnetic storage medium (e.g., a flexible disk, a hard disk, a flash memory, etc.), an optical storage medium (e.g., a CD-ROM, etc.), a magneto-optical storage medium (e.g., an MO, etc.), and the like.

The invention therefore also discloses a storage medium in which a data processing program is stored which is designed to carry out any one of the embodiments of the method according to the invention described above.

It should be understood that although the present description has been described in terms of various embodiments, not every embodiment includes only a single embodiment, and such description is for clarity purposes only, and those skilled in the art will recognize that the embodiments described herein can be combined as a whole to form other embodiments as would be understood by those skilled in the art.

The above description is only a preferred embodiment of the present application and should not be taken as limiting the scope of the present application, and any modifications, equivalents, improvements, etc. made within the spirit and principle of the technical solution of the present application should be included in the scope of the present application.

Claims (10)

1. A method of obtaining a Web browsing interest of a user, applied to a data processing apparatus in which a program having a Web browsing function is installed, comprising:

judging whether the condition of calculating the browsing interest is met, if so, reading a user browsing log from the last time of calculating the browsing interest to the current time;

obtaining interest tags of web pages corresponding to the read user browsing logs and categories to which the interest tags belong;

respectively calculating the weight of the interest tag of each category;

for any category of interest tags, multiplying the last calculated interest calculation value of the category of interest tags by the attenuation factor related to time, then accumulating the obtained value with the weight of the currently calculated interest tags of the category to obtain the interest calculation value of the currently calculated interest tags of the category, storing the currently calculated interest calculation value of the interest tags of the category, and returning to the step of judging whether the condition of calculating browsing interest is met.

2. The method of claim 1, wherein the condition for calculating the browsing interest is: the current system time is a preset time.

3. The method of claim 2, wherein there are a plurality of preset times, the interval between adjacent preset times being at least one day;

the conditions for calculating the browsing interest are as follows: the current system time is one of the preset moments.

4. The method according to claim 3, wherein the interval between adjacent preset times is one day, the browsing interest is calculated for the n +1 th time, and the expression for obtaining the calculated interest value of the current calculation in step E is:

w'(n)＝w(n)·exp(-λ)+ln(PV₀+1)

wherein w' (n) is the calculated interest value on day n +1, and w (n) is the calculated interest value on day n; PV (photovoltaic)₀The browsing times of the web pages of the category interest tags on the (n +1) th day; ln (PV)₀+1) is the weight of the interest tag of the category calculated on day n + 1; λ is a time dependent attenuation factor.

5. The method according to claim 4, characterized in that the time-dependent attenuation factor λ has a value of 0.012.

6. An apparatus for obtaining a Web browsing interest of a user, the apparatus being located at a data processing device in which a program having a Web browsing function is installed, the apparatus comprising: the system comprises a condition judgment module, a reading module, a weight calculation module, an interest calculation module and a storage module;

the condition judgment module is used for judging whether the condition for calculating the browsing interest is met, and if so, the reading module, the weight calculation module and the interest calculation module are enabled;

the reading module is used for reading a user browsing log between the last time of calculating browsing interest and the current time, and acquiring an interest tag of a webpage corresponding to the read user browsing log and a category to which the interest tag belongs;

the weight calculation module is used for calculating the weight of each category of interest tag acquired by the reading module respectively;

the interest calculation module is used for multiplying the interest calculation value of the interest label of the category, which is stored in the storage module at the last time, of the interest label of the category by a time-dependent attenuation factor, then accumulating the interest calculation value of the interest label of the category, which is calculated at the present time, of the weight calculation module to obtain the interest calculation value of the interest label of the category, which is calculated at the present time, and sending the interest calculation value of the interest label of the category, which is calculated at the present time, to the storage module for storage;

the storage module is used for storing the interest calculation value from the interest calculation module.

7. The apparatus of claim 6, wherein the condition for calculating the browsing interest is: the current system time is a preset time.

8. The apparatus of claim 7, wherein there are a plurality of preset times, and the interval between adjacent preset times is at least one day;

the conditions for calculating the browsing interest are as follows: the current system time is one of the preset moments.

9. The apparatus according to claim 8, wherein the interval between adjacent preset times is one day, the browsing interest is calculated for the n +1 th time, and the expression of the calculated interest value obtained by the interest calculating module is as follows:

w'(n)＝w(n)·exp(-λ)+ln(PV₀+1)

wherein w' (n) is the calculated interest value on day n +1, and w (n) is the calculated interest value on day n; PV (photovoltaic)₀The number of views of the web page of the category interest tag on day n +1, ln (PV)₀+1) is the weight of the interest tag of the category calculated on day n + 1; λ is a time dependent attenuation factor.

10. The apparatus of claim 9, wherein the time-dependent attenuation factor λ has a value of 0.012.