A user behavior preference point processing method, system and device

Abstract

The application provides a user behavior preference score processing method, system and device, relates to the technical field of user behavior analysis, and can quickly and accurately obtain an effective user behavior preference score value, so that a user recommendation system quickly and accurately provides a recommendation service for a user. The method comprises the following steps: extracting user behavior preference feature information, wherein the user behavior preference feature information comprises user behavior preference completeness features and user behavior preference persistence features; performing score calculation on the user behavior preference feature information through a user behavior preference score formula with adjustable parameters to obtain a first user behavior preference score value; and sending the first user behavior preference score value to a user recommendation system, wherein the first user behavior preference score value is used to instruct the user recommendation system to determine a recommendation target according to the first user behavior preference score value.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of user behavior analysis, and particularly relates to a user behavior preference score processing method, system and device. BACKGROUND

[0002] User portrait, also known as user role, as an effective tool for sketching target users, associating user demands and design direction, has been widely applied in various fields. In a traditional user portrait system or score system, the score value of a user behavior preference feature is calculated to form a certain user portrait, thereby providing targeted services for users.

[0003] However, in the traditional user portrait system or score system, due to different algorithms for calculating the score value and inconsistent implementation manners, different systems may have different deviations in the final score value, that is, the behavior preference of the same user cannot be expressed as the same score value effect when different algorithms are used to calculate the corresponding score value, and the value significance of different score values cannot be quantitatively compared, thereby causing that when different recommendation systems or machine learning need to be associated and calculated, the score value of the user behavior preference cannot be accurately and quickly calculated, and the effective user behavior preference cannot be quickly and accurately analyzed and expressed, and more accurate recommendation targets of interest for users cannot be provided. SUMMARY

[0004] The embodiments of the present application provide a user behavior preference score processing method, system and device, which can quickly and accurately obtain the effective user behavior preference score value, so that the user recommendation system can quickly and accurately provide recommendation services for users, and the user stickiness and user experience of the user recommendation system are improved.

[0005] In a first aspect, the present application provides a user behavior preference score processing method, comprising: extracting user behavior preference feature information, wherein the user behavior preference feature information comprises user behavior preference completeness feature and user behavior preference persistence feature; performing score calculation on the user behavior preference feature information by using a user behavior preference score formula with adjustable parameters to obtain a first user behavior preference score value; and sending the first user behavior preference score value to a user recommendation system, wherein the first user behavior preference score value is used to instruct the user recommendation system to determine a recommendation target according to the first user behavior preference score value.

[0006] The embodiment of the application is based on the user behavior preference completeness feature and the user behavior preference persistence feature, and the user behavior preference integral formula with adjustable parameters is used for integral calculation, so that the effective first user behavior preference integral value can be quickly and accurately obtained, and the first user behavior preference integral value is sent to the user recommendation system, so that the user recommendation system can quickly and accurately provide the recommendation service for the user, thereby improving the user stickiness and user experience of the user recommendation system.

[0007] In an optional implementation of the first aspect, the adjustable parameters include a score multiplier amplifier a, a completeness offset b, a power parameter m, and an integral offset value k, the integral calculation on the user behavior preference feature information is performed by using the user behavior preference integral formula with adjustable parameters to obtain the first user behavior preference integral value, and the integral calculation includes:

[0008] The values of any one or several of the score multiplier amplifier a, the completeness offset b, the power parameter m, and the integral offset value k are adjusted.

[0009] According to the adjusted values of any one or several of the score multiplier amplifier a, the completeness offset b, the power parameter m, and the integral offset value k, the user behavior preference integral formula with adjustable parameters is updated.

[0010] Integral calculation is performed according to the updated user behavior preference integral formula with adjustable parameters to obtain the first user behavior preference integral value.

[0011] In another optional implementation of the first aspect, the first user behavior preference integral value is sent to the user recommendation system, and the sending includes:

[0012] A user historical behavior preference integral value is obtained.

[0013] A second user behavior preference integral value is determined according to the first user behavior preference integral value and the user historical behavior preference integral value.

[0014] The second user behavior preference integral value is sent to the user recommendation system, and the second user behavior preference integral value is used to instruct the user recommendation system to determine a recommendation target according to the second user behavior preference integral value.

[0015] In another optional implementation of the first aspect, after the second user behavior preference integral value is determined according to the first user behavior preference integral value and the user historical behavior preference integral value, the implementation further includes:

[0016] When the user behavior preference persistence feature meets the decay condition, the second user behavior preference score is reduced to obtain a third user behavior preference score.

[0017] Correspondingly, the second user behavior preference score is sent to the user recommendation system, including:

[0018] The third user behavior preference score is sent to the user recommendation system, and the third user behavior preference score is used to instruct the user recommendation system to determine a recommendation target according to the third user behavior preference score.

[0019] In another optional implementation of the first aspect, when the user behavior preference persistence feature meets the decay condition, the second user behavior preference score is reduced to obtain a third user behavior preference score, including:

[0020] According to the decay period, the second user behavior preference score is reduced by a smoothing decay formula.

[0021] In another optional implementation of the first aspect, when the user behavior preference persistence feature meets the decay condition, the second user behavior preference score is reduced to obtain a third user behavior preference score, including:

[0022] According to the decay period, the second user behavior preference score is reduced by a power decay formula.

[0023] In a second aspect, the application provides a user behavior preference score processing system, including:

[0024] A user behavior preference feature information extraction unit is configured to extract user behavior preference feature information, and the user behavior preference feature information includes user behavior preference completeness features and user behavior preference persistence features.

[0025] A first user behavior preference score calculation unit is configured to calculate a first user behavior preference score by using a user behavior preference score formula with adjustable parameters based on the user behavior preference feature information.

[0026] A first user behavior preference score sending unit is configured to send the first user behavior preference score to a user recommendation system, and the first user behavior preference score is used to instruct the user recommendation system to determine a recommendation target according to the first user behavior preference score.

[0027] In an optional implementation of the second aspect, the adjustable parameters include a score multiplier a, a completeness offset b, a power parameter m, and an integral offset k, and the first user behavior preference score calculation unit includes:

[0028] a parameter adjustment subunit configured to adjust a value of any one or more of the score multiplier a, the completeness offset b, the power parameter m, and the integral offset k;

[0029] a formula updating subunit configured to update the user behavior preference score formula with adjustable parameters according to the adjusted value of any one or more of the score multiplier a, the completeness offset b, the power parameter m, and the integral offset k;

[0030] a first user behavior preference score calculation subunit configured to perform integral calculation according to the updated user behavior preference score formula with adjustable parameters to obtain the first user behavior preference score.

[0031] In another optional implementation of the second aspect, the first user behavior preference score sending unit includes:

[0032] a user historical behavior preference score obtaining subunit configured to obtain a user historical behavior preference score;

[0033] a second user behavior preference score calculation subunit configured to determine a second user behavior preference score according to the first user behavior preference score and the user historical behavior preference score;

[0034] a second user behavior preference score sending subunit configured to send the second user behavior preference score to a user recommendation system, where the second user behavior preference score is used to instruct the user recommendation system to determine a recommendation target according to the second user behavior preference score.

[0035] In another optional implementation of the second aspect, the first user behavior preference score sending unit further includes:

[0036] a third user behavior preference score calculation subunit configured to perform value reduction processing on the second user behavior preference score to obtain a third user behavior preference score when a user behavior preference persistence feature meets a decay condition;

[0037] a third user behavior preference score sending subunit configured to send the third user behavior preference score to a user recommendation system, where the third user behavior preference score is used to instruct the user recommendation system to determine a recommendation target according to the third user behavior preference score.

[0038] In another optional implementation of the second aspect, the third user behavior preference score calculation subunit is configured to:

[0039] According to the decay period, the second user behavior preference score is processed by a smoothing decay formula.

[0040] In another optional implementation of the second aspect, the third user behavior preference score calculation subunit is configured to:

[0041] According to the decay period, the second user behavior preference score is processed by a power decay formula.

[0042] In a third aspect, the present application provides a user behavior preference score processing device, comprising a processor, a memory, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the method of the first aspect or any optional implementation of the first aspect.

[0043] In a fourth aspect, the present application provides a computer readable storage medium, which stores a computer program, wherein the computer program is executed by a processor to implement the method of the first aspect or any optional implementation of the first aspect.

[0044] In a fifth aspect, the present application provides a computer program product, which, when executed on a user behavior preference score processing device, causes the user behavior preference score processing device to perform the steps of the user behavior preference score processing method of the first aspect.

[0045] It can be understood that the beneficial effects of the second aspect to the fifth aspect can be referred to the related description in the first aspect, which will not be repeated here. BRIEF DESCRIPTION OF DRAWINGS

[0046] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description are only some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained without creative labor.

[0047] Figure 1 is a flowchart of a user behavior preference score processing method provided by the embodiments of the present application;

[0048] Figure 2 is a flowchart of another user behavior preference score processing method provided by the embodiments of the present application;

[0049] Figure 3is a structural schematic diagram of a user behavior preference point processing system provided by an embodiment of the present application.

[0050] Figure 4 is a structural schematic diagram of a user behavior preference point processing device provided by an embodiment of the present application. DETAILED DESCRIPTION

[0051] In the following description, specific details are set forth, such as a particular system architecture, techniques, etc., in order to provide a thorough understanding of the embodiments of the present application. However, persons skilled in the art will understand that the present application can be practiced without these specific details. In other instances, well-known systems, structures, circuits, and techniques have not been shown in detail in order not to obscure the understanding of this description.

[0052] It should be understood that the term "and / or" used in the description of the present application and the appended claims means one or more of the associated listed items as well as all possible combinations of the items and includes these combinations. In addition, in the description of the present application and the appended claims, the terms "first", "second", "third", etc. are only used to distinguish the description and cannot be understood as indicating or implying relative importance.

[0053] It should also be understood that the reference "one embodiment" or "some embodiments" and the like described in the present application means that a particular feature, structure or characteristic described in connection with the embodiment is included in at least one embodiment of the present application. Thus, the appearance of the phrases "in one embodiment", "in some embodiments", "in other some embodiments", "in yet some embodiments" and the like in various places in the specification is not necessarily all referring to the same embodiment, but means "one or more but not all embodiments", unless otherwise specifically stated. The terms "include", "contain", "have" and their variants mean "including but not limited to", unless otherwise specifically stated.

[0054] In the embodiments of the present application, the user portrait system or other point system that needs to perform point calculation is collectively referred to as a point system. When the present application performs point association calculation in two different point systems, the point value of the user behavior preference can be quickly and accurately calculated, the effective user behavior preference can be quickly and accurately expressed, more interesting recommended targets can be provided for the user, the effect of quick response is achieved, and the user experience is improved.

[0055] Please refer to Figure 1 , Figure 1 is a flowchart of a user behavior preference point processing method provided by an embodiment of the present application, which is described in detail as follows:

[0056] In step S101, user behavior preference feature information is extracted, which includes user behavior preference completeness feature and user behavior preference persistence feature.

[0057] In the embodiments of the present application, user behavior preference information is extracted from user behavior information. In particular, in order to better analyze user behavior preference, user behavior information in one or more predetermined periods is acquired, and user behavior preference information is extracted from the user behavior information in the one or more predetermined periods. The predetermined period is a time period, such as one hour, one day, one week, two weeks, etc. The predetermined period is not limited herein.

[0058] In order to facilitate quantification of user behavior preference feature information, different preset values can be used to quantize different user behavior preference features, so as to facilitate calculation of the score of user behavior preference.

[0059] For example, the user behavior preference completeness feature is represented by a first value between 0 and 1.0, which is used to infer the completeness of user behavior preference. For example, when a user watches a program completely in a predetermined period, 1.0 is used to represent that the completeness of user behavior preference is 1; when a user watches a program for 50% in a predetermined period, 0.5 is used to represent that the completeness of user behavior preference is 0.5; when a user watches a program for 10% in a predetermined period, 0.1 is used to represent that the completeness of user behavior preference is 0.1, and so on.

[0060] The user behavior preference persistence feature is represented by a second value, which is a positive integer. The second value is used to represent the persistence of user behavior preference. For example, the user behavior preference persistence feature in a single period is represented by 1, which represents that the persistence of user behavior preference is 1; the user behavior preference persistence feature in two consecutive periods is represented by 2, which represents that the persistence of user behavior preference is 2, and so on.

[0061] It should be noted that when the persistence of user behavior preference is greater than 1, i.e., the user behavior preference persistence feature is a value greater than 1, the user behavior preference completeness feature in the period corresponding to the user behavior preference persistence feature is accumulated. For example, the persistence of user behavior preference is 3, i.e., the user has the same behavior preference in three consecutive periods, and the completeness of user behavior preference corresponding to each period is 0.9, 1.0, and 0.2. Therefore, the completeness of user behavior preference in the three periods is 2.1, i.e., the sum of 0.9, 1.0, and 0.2.

[0062] Step S102, the user behavior preference characteristic information is integrated by a user behavior preference integral formula with adjustable parameters to obtain a first user behavior preference integral value.

[0063] In the embodiments of the present application, the user behavior preference integral formula with adjustable parameters can be adjusted by adjusting the values of the adjustable parameters, wherein the adjustable parameters include a score rate amplifier a, a completeness offset b, a power number adjustment value m, and an integral offset value k. Any one or several of the score rate amplifier a, the completeness offset b, the power number adjustment value m, and the integral offset value k are adjusted. Then, the user behavior preference integral formula with adjustable parameters is updated according to any one or several of the adjusted score rate amplifier a, the completeness offset b, the power number adjustment value m, and the integral offset value k. The user behavior preference completeness characteristic and the user behavior preference persistence characteristic are quantitatively calculated according to the updated user behavior preference integral formula with adjustable parameters, and the integral value of the user behavior preference, i.e., the first user behavior preference integral value, can be obtained.

[0064] Most of the current systems for calculating the integral value of user behavior preference generally use linear calculation. Different user behavior preference integrals are calculated by a linear integral formula. However, due to the characteristics of linear calculation, the integral value of user behavior preference cannot be quickly improved by using the linear integral calculation formula, and the effective user behavior preference cannot be quickly determined. In order to quickly and efficiently represent the user behavior preference characteristic, the power calculation is used in the behavior preference integral formula with adjustable parameters, which can quickly improve the integral value of the effective user behavior preference and relatively reduce the integral value of the ineffective user behavior preference, so as to ensure that the integral value of the user behavior preference can more efficiently and accurately express the user behavior preference.

[0065] It should be noted that the effective user behavior preference refers to the user behavior preference whose sum of the completeness and the completeness offset is greater than or equal to 1, i.e., the user behavior preference whose completeness after adding the completeness offset is greater than or equal to 1.

[0066] Value

[0067] Here, the user behavior preference integral formula with adjustable parameters is:

[0068] F=a*(X+b) (Y+m) +k (1)

[0069] Wherein, F represents the integral value of user behavior preference; X represents the user behavior preference completeness characteristic; Y represents the user behavior preference persistence characteristic; a represents the score rate amplifier; b represents the completeness offset; m represents the power number adjustment value; and k represents the integral offset value.

[0070] Wherein, the score multiplier amplifier a is greater than 0, and the default value is 1; the completeness offset b, whose default value is 0.2, has a value range of 0-1; the power adjustment value m, whose default value is 1, usually has a value range greater than 1; the integral offset value k, whose default value is 0, usually has a value range greater than or equal to 0.

[0071] It should be noted that the adjustable parameter is a parameter that can be adjusted, such as a, b, m, and k in formula (1). When the two different integral systems are associated and calculated, such as a user portrait system or an integral system, the two different integral systems can be generally comparable in speed or comparability by adjusting the adjustable parameter.

[0072] In general, the value of a is 1, the value of m is 1, and the value of k is 0.

[0073] It should be noted that the value of a is related to the integral maximum value of the two integral systems associated and calculated. When the integral maximum values of the two integral systems are the same, the value of a is 1. When the integral maximum values of the two integral systems are different, the value of a is the ratio of the integral maximum values of the two integral systems. For example, the integral maximum value of integral system A is 200 points, and the integral maximum value of integral system B is 400 points. When integral system B needs to call the user behavior preference integral value from integral system A, the value of a needs to be adjusted to 2, i.e. 400 ÷ 200 = 2, to ensure that the behavior characteristics of each integral system are consistent.

[0074] In some embodiments of the present application, when the user behavior preference integral value needs to be calculated quickly, or the integral maximum values of the two compared integral systems are not equal, such as one integral maximum value is 500 points and the other integral maximum value is 1000 points, the score multiplier amplifier a in the behavior preference integral formula of the latter can be 2, so that the user behavior preference integral values of the two integral systems are generally comparable in calculation speed or comparability. The values of m and k can also be appropriately increased for adjustment. The following examples are illustrated by combining typical period data, where a = 1 and b = 0.5.

[0075] When the duration Y of the user's behavior preference is 1, i.e. the user behavior preference persistence characteristic is 1, the typical value table is shown in Table 1:

[0076] Completeness X Duration Y m＝1 k＝0 Integral value 0.1 1 1 0 0.36 0.3 1 1 0 0.64 0.5 1 1 0 1.0 0.7 1 1 0 1.44 0.9 1 1 0 1.96 1 1 1 0 2.25

[0077] Table 1

[0078] As can be seen from Table 1, under the condition that the duration of the user's behavior preference is 1, when the completeness X of the user's behavior preference is less than 0.5, the integral value is not more than 1 point; when the completeness is greater than 0.5, the integral value is between 1-2.25.

[0079] When the persistence Y of the user's behavior preference is 2, i.e. the persistence characteristic of the user's behavior preference is 2, the typical value table is shown in Table 2:

[0080] Completeness X Duration Y m＝1 k＝0 Integral value 0.5 2 1 0 1 1.0 2 1 0 3.38 1.5 2 1 0 8 2 2 1 0 15.6

[0081] Table 2

[0082] From Table 2, it can be seen that when the persistence Y of the user's behavior preference is 2, if the completeness X is less than 0.5, the integral value is still not more than 1; if the completeness is greater than 0.5, the integral value is between 1 and 15.6. If the completeness exceeds 75% for two consecutive periods, i.e. 0.75*2=1.5, the integral value is greater than 8, the integral value is obviously improved, i.e. under the condition of the same persistence Y, the higher the completeness X, the more obvious the integral value is, for example, X*Y=4 is obviously much larger than the integral value calculated by using the general linear integral calculation formula, and is improved by nearly 4 times.

[0083] When the persistence Y of the user's behavior preference is 3, i.e. the persistence characteristic of the user's behavior preference is 3, the typical value table is shown in Table 3:

[0084] Completeness X Duration Y m＝1 k＝0 Integral value 0.5 3 1 0 1 1.0 3 1 0 5.06 1.5 3 1 0 16 2 3 1 0 39.06 2.25 3 1 0 57.19 3 3 1 0 150.06

[0085] Table 3

[0086] From Table 3, it can be seen that when the persistence Y of the user's behavior preference is 3, if the completeness X of the user's behavior preference is less than 0.5, the integral value is still not more than 1; if the completeness is greater than 0.5, the integral value is between 1 and 150.06. If the completeness exceeds 75% for three consecutive periods, i.e. 0.75*3=2.25, the integral value is 57.19, the integral value is obviously improved, for example, the integral value 150.06 calculated by using the user behavior preference integral formula with adjustable parameters in the last row of Table 3 is obviously much larger than the integral value calculated by using the general linear integral calculation formula, i.e. X*Y=9, and is improved by more than 16 times. Obviously, the higher the completeness of the user's behavior preference and the higher the persistence, the more obvious the integral value is improved, and the effective user behavior preference can be quickly expressed by the integral value, so that the target of interest of the user can be grasped, and more recommended targets of interest of the user can be provided.

[0087] For the typical value table when the persistence Y of the user's behavior preference is greater than 3, no example is given here. In Tables 1-3, the integral value is a short name of the user behavior preference integral value.

[0088] It can be inferred that for the continuous multi-period repeated user behavior preference, i.e. the user behavior preference persistence feature repeats the same and the completeness is high, the first user behavior preference integral value calculated by formula (1) obviously presents a power level effect, which can quickly lock the user's behavior preference and has a very important indication effect for a real-time recommendation system.

[0089] It should be noted that the purpose of parameter adjustment is to adjust according to the use scene of the specific user behavior preference integral value. For example, when it is determined that a user behavior preference that is repeated for 3 consecutive periods, i.e. the persistence is 3 and the completeness is good, represents a user behavior preference of "very like", the recommendation system or other system determines the threshold value of this user behavior feature of "very like" and applies it. Taking the above table as an example, the current user behavior preference integral value can reach 150.06 points, and 75% is taken as the threshold value percentage of "very like", then 150.06 ÷ 75% = 200.08 can be calculated. Therefore, it can be determined that the integral maximum value of the integral system is 200 points, which is very suitable. If the integral maximum value of the other integral system is 1000 points, then correspondingly, in order to keep the behavior feature determination of each integral system consistent, the integral maximum value of the integral system of 200 points needs to be enlarged by 5 times, i.e. the parameter a in the user behavior preference integral value formula needs to be adjusted to 5.

[0090] In some embodiments of the present application, the user behavior preference integral formula contains rewards for valid user behavior preferences and penalties for invalid user behavior preferences, and the value range of valid user behavior preferences can be determined by adjusting the completeness offset b appropriately.

[0091] In an actual application scene, such as the user's video watching behavior, because the video generally includes the beginning and end of the drama, the user generally does not watch the entire video content at 100%, so when the user watches the video content to 90% or more, it can be considered that the user has completed the watching behavior of the video content at 100%. Therefore, in actual operation, it is necessary to normalize the completeness of less than 90% of the watching behavior. That is:

[0092] The normalized completeness of the watching behavior = the completeness of less than 90% ÷ 0.9

[0093] When the completeness of the watching behavior is greater than or equal to 90%, the completeness of the watching behavior is 1.

[0094] It should be noted that in the user behavior preference integral formula with adjustable parameters, each adjustable parameter has its practical significance, and in the embodiments of the present application, only a typical parameter value table is cited, which is only a part of the expression of the formula effect.

[0095] Step S103, sending the first user behavior preference integral value to a user recommendation system, wherein the first user behavior preference integral value is used to instruct the user recommendation system to determine a recommendation target according to the first user behavior preference integral value.

[0096] In the embodiments of the present application, the recommendation target includes but is not limited to movies, goods, music, etc. After the user behavior preference integral value is quickly and accurately calculated by the user behavior preference integral formula with adjustable parameters, the user behavior preference integral value is sent to the user recommendation system, which can help the user recommendation system to recommend the target preferred or interested by the user more quickly and accurately, thereby improving the user stickiness and use experience.

[0097] Please refer to Figure 2 , Figure 2 is another flowchart of a user behavior preference integral processing method provided by the embodiments of the present application. Figure 2 The method shown in Figure 1 On the basis of the method shown in

[0098] Step S201, obtaining a user historical behavior preference integral value.

[0099] In the embodiments of the present application, the user historical behavior preference integral value is the user behavior preference integral value recorded in the last period.

[0100] It should be noted that the user historical behavior preference integral value is the sum of the user behavior preference integral values calculated in each period before the last period and the user behavior preference integral value calculated in the last period, that is, the user historical behavior preference integral value is the sum of the user behavior preference integral values calculated in multiple periods before the current period.

[0101] Step S202, determining a second user behavior preference integral value according to the first user behavior preference integral value and the user historical behavior preference integral value.

[0102] In the embodiments of the present application, the second user behavior preference integral value is the sum of the first user behavior preference integral value and the user historical behavior preference integral value, that is, the user historical behavior preference integral value is added to the first user behavior preference integral value, and the finally added user behavior preference integral value is used as a reference value provided to the user system.

[0103] Step S203, sending the second user behavior preference integral value to a user recommendation system, wherein the second user behavior preference integral value is used to instruct the user recommendation system to determine a recommendation target according to the second user behavior preference integral value.

[0104] In the embodiments of the present application, the second user behavior preference score obtained after accumulation is sent to the user recommendation system, so that the user recommendation system can more accurately recommend services for the user according to the second user behavior preference score.

[0105] In order to further improve the accuracy of the user behavior preference score and reduce the influence of invalid user behavior preference on the score, when the user behavior preference persistence feature does not appear for a plurality of consecutive periods, such as two or more consecutive periods, the user behavior preference score should enter a decay period, and the calculated user behavior preference score is reduced by a decay mechanism.

[0106] When the user behavior preference persistence feature meets the decay condition, the second user behavior preference score is reduced to obtain a third user behavior preference score, and the third user behavior preference score is sent to the user recommendation system. The third user behavior preference score is used to indicate that the user recommendation system determines the recommendation target according to the third user behavior preference score.

[0107] Here, the decay condition is that the user behavior preference persistence feature does not appear for two or more consecutive periods, that is, the user behavior preference persistence feature appears 0 for two or more times, that is, the user behavior preference persistence feature meets the decay condition.

[0108] In some embodiments of the present application, the way of reducing the second user behavior preference score can adopt linear reduction or power level reduction, so that the user's behavior feature quickly drops below the threshold required by the user recommendation system.

[0109] In a specific embodiment of the present application, the second user behavior preference score is reduced by a smoothing decay formula according to the decay period. The smoothing decay formula is:

[0110] F=F*(1-0.1*y)

[0111] Wherein, F is the user behavior preference score, and y is the decay period.

[0112] In the smoothing decay formula, the user behavior preference score after decay is the product of the user behavior preference score before decay and the difference between 1 and 0.1 multiplied by the decay period.

[0113] In a specific embodiment of the present application, the second user behavior preference score is reduced by a power decay formula according to the decay period. The power decay formula is:

[0114] F=F*(1-0.1*(d y ))

[0115] Wherein, F is the user behavior preference integral value, d is the power level coefficient, and y is the decay period. Here, d = 2.

[0116] In one application scenario, taking 1000 as the maximum value of the user behavior preference integral value as an example, when there is no user behavior preference persistence feature for two or more consecutive periods, the above two decay formulas are used to reduce the user behavior preference integral value, and the typical value table obtained is shown in Table 4:

[0117] As can be seen from Table 4, if 66.7% (667 points) or more is very like, 33.3% (333 points) or more is like, and 33.3% or less is attention state, then according to the calculation result of the above typical value table, the smoothing decay needs to be decayed to the like state range in the 3rd period and to the attention state in the 5th period; and the power decay only needs to be decayed to the like state in the 2nd period and to the attention state in the 3rd period. As can be seen, the decay speed of the power decay is very fast, only 4 gradients, and the smoothing decay has 9 gradients.

[0118] As can be seen, through the decay formula, especially the power decay formula, the user behavior preference integral value can be quickly reduced, the invalid user behavior preference can be quickly reduced, the invalid user behavior preference integral value can be pulled away or adjusted to the interval value range outside the user recommendation system that does not need to be kept attention, so as to quickly and efficiently adjust the data output of the user recommendation system and achieve the effect of quick response.

[0119] In the embodiments of the present application, based on the user behavior preference integrity feature and the user behavior preference persistence feature, the user behavior preference integral formula with adjustable parameters is used for integral calculation, the effective first user behavior preference integral value can be quickly and accurately obtained, and the first user behavior preference integral value is sent to the user recommendation system, so that the user recommendation system can quickly and accurately provide the recommendation service for the user, thereby improving the user stickiness and user experience of the user recommendation system.

[0120] It should be understood that the size of the serial number of each step in the above embodiments does not mean the order of execution, and the execution order of each process should be determined according to its function and inherent logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0121] Based on the user behavior preference integral processing method provided in the above embodiments, the system embodiment for implementing the above method embodiment is further provided in the embodiments of the present application.

[0122] Please refer to Figure 3 , Figure 3is a schematic diagram of a user behavior preference score processing system provided by the embodiment of the application. Each unit included is used to execute Figure 1 each step in the corresponding embodiment. For details, please refer to Figure 1 the relevant description in the corresponding embodiment. For ease of illustration, only the part related to the embodiment is shown. Please refer to Figure 3 , the user behavior preference score processing system 3 includes:

[0123] The user behavior preference feature information extraction unit 31 is configured to extract user behavior preference feature information, wherein the user behavior preference feature information includes user behavior preference completeness characteristics and user behavior preference persistence characteristics.

[0124] The first user behavior preference score calculation unit 32 is configured to calculate a first user behavior preference score by using a user behavior preference score formula with adjustable parameters based on the user behavior preference feature information.

[0125] The first user behavior preference score sending unit 33 is configured to send the first user behavior preference score to a user recommendation system, wherein the first user behavior preference score is used to instruct the user recommendation system to determine a recommendation target based on the first user behavior preference score.

[0126] In the embodiment of the application, the adjustable parameters include a score multiplier amplifier a, a completeness offset b, a power number adjustment value m, and an integral offset value k. The first user behavior preference score calculation unit 32 includes:

[0127] The parameter adjustment subunit is configured to adjust the value of any one or several of the score multiplier amplifier a, the completeness offset b, the power number adjustment value m, and the integral offset value k.

[0128] The formula updating subunit is configured to update the user behavior preference score formula with adjustable parameters based on the adjusted value of any one or several of the score multiplier amplifier a, the completeness offset b, the power number adjustment value m, and the integral offset value k.

[0129] The first user behavior preference score calculation subunit is configured to calculate the first user behavior preference score based on the updated user behavior preference score formula with adjustable parameters.

[0130] In some embodiments of the application, the first user behavior preference score sending unit 33 includes:

[0131] The user historical behavior preference score obtaining subunit is configured to obtain a user historical behavior preference score.

[0132] The second user behavior preference score value calculation sub-unit is configured to determine a second user behavior preference score value according to the first user behavior preference score value and the user historical behavior preference score value.

[0133] The second user behavior preference score value sending sub-unit is configured to send the second user behavior preference score value to a user recommendation system, where the second user behavior preference score value is used to instruct the user recommendation system to determine a recommendation target according to the second user behavior preference score value.

[0134] In some embodiments of the present application, the first user behavior preference score value sending unit 33 further includes:

[0135] The third user behavior preference score value calculation sub-unit is configured to perform a value reduction process on the second user behavior preference score value to obtain a third user behavior preference score value when the user behavior preference persistence feature satisfies a decay condition.

[0136] The third user behavior preference score value sending sub-unit is configured to send the third user behavior preference score value to a user recommendation system, where the third user behavior preference score value is used to instruct the user recommendation system to determine a recommendation target according to the third user behavior preference score value.

[0137] In some embodiments of the present application, the third user behavior preference score value calculation sub-unit is specifically configured to:

[0138] The second user behavior preference score value is reduced by a smoothing decay formula according to a decay period.

[0139] In some embodiments of the present application, the third user behavior preference score value calculation sub-unit is specifically configured to:

[0140] The second user behavior preference score value is reduced by a power decay formula according to a decay period.

[0141] It should be noted that the information interaction, execution process and the like between the above modules are based on the same concept as the method embodiments of the present application, and the specific functions and technical effects brought by the same can be referred to the method embodiments part, and will not be described here.

[0142] Figure 4 is a schematic diagram of a user behavior preference score processing device provided by an embodiment of the present application. As shown in Figure 4As shown, the user behavior preference score processing device 4 of this embodiment comprises a processor 40, a memory 41, and a computer program 42 stored in the memory 41 and executable on the processor 40, such as a speech recognition program. The processor 40 implements the steps in each of the above user behavior preference score processing method embodiments when executing the computer program 42, such as Figure 1 As shown, the steps 101-103. Alternatively, the processor 40 implements the functions of each module / unit in each of the above system embodiments when executing the computer program 42, such as Figure 3 As shown, the functions of the units 31-33.

[0143] For example, the computer program 42 can be divided into one or more modules / units, one or more modules / units are stored in the memory 41 and executed by the processor 40 to complete the present application. One or more modules / units can be a series of computer program instruction segments capable of completing a specific function, which is used to describe the execution process of the computer program 42 in the user behavior preference score processing device 4. For example, the computer program 42 can be divided into a user behavior preference feature information extraction unit 31, a first user behavior preference score value calculation unit 32, and a first user behavior preference score value sending unit 33, and the specific functions of each unit are described in Figure 1 Corresponding description in the embodiments, which is not described here.

[0144] The user behavior preference score processing device can include, but is not limited to, the processor 40, the memory 41. Those skilled in the art can understand that, Figure 4 The user behavior preference score processing device 4 is only an example and does not constitute a limitation on the user behavior preference score processing device 4, and can include more or fewer components than shown, or combine certain components, or different components, such as the user behavior preference score processing device can also include input / output devices, network access devices, buses, etc.

[0145] The processor 40 can be a central processing unit (CPU), and can also be other general-purpose processors, digital signal processors (DSP), application specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic components, discrete hardware components, etc. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor.

[0146] The memory 41 can be an internal storage unit of the user behavior preference point processing device 4, for example, a hard disk or a memory of the user behavior preference point processing device 4. The memory 41 can also be an external storage device of the user behavior preference point processing device 4, for example, a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. equipped on the user behavior preference point processing device 4. Further, the memory 41 can also include both the internal storage unit and the external storage device of the user behavior preference point processing device 4. The memory 41 is used to store computer programs and other programs and data required by the user behavior preference point processing device. The memory 41 can also be used to temporarily store data that has been output or will be output.

[0147] The embodiment of the present application further provides a computer readable storage medium, which stores a computer program. The computer program is executed by a processor to implement the user behavior preference point processing method.

[0148] The embodiment of the present application provides a computer program product, which, when running on a user behavior preference point processing device, enables the user behavior preference point processing device to implement the user behavior preference point processing method.

[0149] Those skilled in the art can clearly understand that, for the convenience and brevity of description, only the division of the above functional units and modules is exemplified, and in actual application, the above functions can be completed by different functional units and modules according to needs, that is, the internal structure of the system is divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated in one processing unit, or each unit can be physically present separately, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware or in the form of software functional unit. In addition, the specific names of the functional units and modules are only for easy distinction, and do not limit the protection scope of the present application. The specific working process of the units and modules in the system can refer to the corresponding process in the foregoing method embodiments, which will not be described here.

[0150] In the above embodiments, the description of each embodiment has its own emphasis, and the parts not described or recorded in detail in a certain embodiment can be referred to the related description of other embodiments.

[0151] Those skilled in the art can understand that the units and algorithm steps of each example described in combination with the embodiments disclosed herein can be realized in electronic hardware or a combination of computer software and electronic hardware. Whether the functions are realized in hardware or software manner depends on the specific application and design constraints of the technical solution. The skilled person can use different methods to realize the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0152] The above-described embodiments are only used to illustrate the technical solutions of the present application, but not limit the present application; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that the technical solutions recorded in the foregoing embodiments can be modified, or some technical features can be replaced by equivalent replacements; 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 embodiments of the present application, and should be included in the protection scope of the present application.

Claims

1. A user behavior preference point processing method, characterized by, The method comprises: extracting user behavior preference feature information, the user behavior preference feature information comprising a user behavior preference completeness feature and a user behavior preference persistence feature; wherein the user behavior preference completeness feature is the degree to which a user watches a program within a predetermined period, and the user behavior preference persistence feature is the number of periods in which the user behavior persists; integrating the user behavior preference feature information by using a user behavior preference integration formula with adjustable parameters to obtain a first user behavior preference integration value; sending the first user behavior preference integration value to a user recommendation system, the first user behavior preference integration value being used to instruct the user recommendation system to determine a recommendation target based on the first user behavior preference integration value; wherein the user behavior preference integration formula with adjustable parameters is: wherein, represents a user behavior preference score; represents a user behavior preference completeness characteristic; represents a user behavior preference persistence characteristic; represents a score multiplier amplifier; represents a completeness offset; represents a power number tuning value; represents an offset value for the score; the integration of the user behavior preference feature information by using the user behavior preference integration formula with adjustable parameters to obtain a first user behavior preference integration value comprises: Adjust the fraction amplifier The integrity offset The power parameter tuning value and the integral offset value any one or more values ​​in the range; According to the adjusted fractional amplifier The integrity offset The power parameter tuning value and the integral offset value Update the user behavior preference integral formula with adjustable parameters using any one or more of the parameters; integrating the user behavior preference feature information according to the updated user behavior preference integration formula with adjustable parameters to obtain the first user behavior preference integration value.

2. The user behavior preference point processing method of claim 1, wherein, the sending of the first user behavior preference integration value to a user recommendation system comprises: obtaining a user historical behavior preference integration value; determining a second user behavior preference integration value based on the first user behavior preference integration value and the user historical behavior preference integration value; sending the second user behavior preference integration value to a user recommendation system, the second user behavior preference integration value being used to instruct the user recommendation system to determine a recommendation target based on the second user behavior preference integration value.

3. The user behavior preference point processing method of claim 2, wherein, after the determination of the second user behavior preference integration value based on the first user behavior preference integration value and the user historical behavior preference integration value, the method further comprises: when the user behavior preference persistence feature satisfies a decay condition, performing a value reduction process on the second user behavior preference integration value to obtain a third user behavior preference integration value; correspondingly, the sending of the second user behavior preference integration value to a user recommendation system comprises: sending the third user behavior preference integration value to a user recommendation system, the third user behavior preference integration value being used to instruct the user recommendation system to determine a recommendation target based on the third user behavior preference integration value.

4. The user behavior preference point processing method of claim 3, wherein, the performing of the value reduction process on the second user behavior preference integration value to obtain a third user behavior preference integration value when the user behavior preference persistence feature satisfies a decay condition comprises: performing a value reduction process on the second user behavior preference integration value by using a smooth decay formula according to a decay period.

5. The user behavior preference point processing method of claim 3, wherein, the performing of the value reduction process on the second user behavior preference integration value to obtain a third user behavior preference integration value when the user behavior preference persistence feature satisfies a decay condition comprises: performing a value reduction process on the second user behavior preference integration value by using a power decay formula according to a decay period.

6. A user behavior preference points processing system characterized by, the system comprises: The user behavior preference feature information extraction unit is configured to extract user behavior preference feature information, which includes user behavior preference completeness feature and user behavior preference persistence feature. The user behavior preference completeness feature is the degree of watching a program by a user in a predetermined period, and the user behavior preference persistence feature is the number of periods of user behavior persistence. The first user behavior preference score calculation unit is configured to calculate a first user behavior preference score by using a user behavior preference score formula with adjustable parameters based on the user behavior preference feature information. The first user behavior preference score sending unit is configured to send the first user behavior preference score to a user recommendation system, which is used to instruct the user recommendation system to determine a recommendation target based on the first user behavior preference score. The user behavior preference score formula with adjustable parameters is as follows: wherein, represents a user behavior preference score; represents a user behavior preference completeness characteristic; represents a user behavior preference persistence characteristic; represents a score multiplier amplifier; represents a completeness offset; represents a power number tuning value; represents an offset value for the score; The first user behavior preference score calculation unit includes: The parameter adjustment subunit is used to adjust the fraction amplifier. The integrity offset The power parameter tuning value and the integral offset value any one or more values ​​in the range; a formula updating subunit configured to update the user behavior preference integral formula with adjustable parameters according to any one or several of the following: the adjusted score multiplier amplifier , the completeness offset , the power adjustment value , and the integral offset value ​ The first user behavior preference score calculation subunit is configured to calculate the first user behavior preference score based on the user behavior preference score formula with adjustable parameters.

7. A user behavior preference point processing apparatus, characterized by, The computer program is executed by the processor to implement the user behavior preference score processing method according to any one of claims 1 to 5.

8. A computer-readable storage medium storing a computer program, the computer-readable storage medium comprising: The computer program is executed by the processor to implement the user behavior preference score processing method according to any one of claims 1 to 5.

Images