Information recommendation method, device and electronic device of recommendation system

Abstract

The embodiments of the present application provide an information recommendation method, device, and electronic device for a recommendation system, which obtains the target data of the target user and determines the user level to which the target user belongs based on the target data. According to the user level, the targets of the recall module and the sorting module included in the target object in the recommendation system are adjusted to the same collaborative target. The target information is determined using the collaborative target, and the target information is recommended to the target user. Therefore, the collaborative targets of the recall module and the sorting module are determined to be consistent, so as to achieve the purpose of collaboration between the recall module and the sorting module, and the target information in the recommendation list is also more likely to meet the user's information needs, thereby improving the user experience.

Description

Technical Field

[0001] The present invention relates to the field of information processing technology, and in particular to an information recommendation method, device and electronic equipment of a recommendation system. Background Art

[0002] The recommendation system recommends information of interest to users based on their information needs and interests, thereby guiding users to discover their own information needs.

[0003] In some scenarios, the recommendation system includes a recall module and a sorting module. The recall module has multiple recall channels, which can cooperate with each other. The sorting module also has multiple sorting channels, which can cooperate with each other to determine the recommendation list. However, the lack of coordination between the recall module and the sorting module results in that the information in the recommendation list cannot meet the user's information needs, resulting in a low user experience. Summary of the invention

[0004] The purpose of the embodiments of the present application is to provide an information recommendation method, device and electronic device for a recommendation system, so as to coordinate between a recall module and a sorting module so that the information in the recommendation list meets the information needs of the user and improves the user experience.

[0005] In order to solve the above technical problems, the embodiments of the present application are implemented as follows:

[0006] In a first aspect, an embodiment of the present application provides an information recommendation method of a recommendation system, the method comprising:

[0007] Obtain target data of a target user; determine the user level to which the target user belongs based on the target data; adjust the targets of a recall module and a sorting module included in a target object in a recommendation system to the same collaborative target based on the user level; determine target information using the collaborative target, and recommend the target information to the target user.

[0008] In a second aspect, an embodiment of the present application provides an information recommendation device of a recommendation system, the device comprising:

[0009] An acquisition module is used to acquire target data of a target user; a first determination module is used to determine the user level to which the target user belongs based on the target data; a second determination module is used to adjust the targets of a recall module and a sorting module included in the target object in the recommendation system to the same collaborative target based on the user level; a recommendation module is used to determine target information using the collaborative target and recommend the target information to the target user.

[0010] In a third aspect, an embodiment of the present application provides an electronic device, comprising a processor, a communication interface, a memory and a communication bus; wherein the processor, the communication interface and the memory communicate with each other through the bus; the memory is used to store computer programs; the processor is used to execute the programs stored in the memory to implement the information recommendation method steps of the recommendation system as described in the first aspect.

[0011] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the information recommendation method of the recommendation system described in the first aspect are implemented.

[0012] In a fifth aspect, an embodiment of the present application provides a chip, comprising a processor and a communication interface, wherein the communication interface is coupled to the processor, and the processor is used to run a program or instruction to implement the information recommendation method of the recommendation system as described in the first aspect.

[0013] It can be seen from the technical solution provided by the above embodiments of the present application that by obtaining the target data of the target user, the user level to which the target user belongs is determined based on the target data. According to the user level, the goals of the recall module and the sorting module included in the target object in the recommendation system are adjusted to the same collaborative goal. The collaborative goal is used to determine the target information, and the target information is recommended to the target user. Therefore, the collaborative goals of the recall module and the sorting module are determined to be consistent, so as to achieve the purpose of collaboration between the recall module and the sorting module, and the target information in the recommendation list is also more likely to meet the user's information needs, thereby improving the user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0015] Figure 1 A first flow chart of an information recommendation method of a recommendation system is provided for an embodiment of the present application;

[0016] Figure 2 A second flow chart of an information recommendation method of a recommendation system is provided for an embodiment of the present application;

[0017] Figure 3 A third flow chart of an information recommendation method of a recommendation system is provided for an embodiment of the present application;

[0018] Figure 4 A flowchart of an information recommendation method of a recommendation system is provided for an embodiment of the present application;

[0019] Figure 5 A schematic diagram of a functional module for information recommendation of a recommendation system is provided for an embodiment of the present application;

[0020] Figure 6 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0021] The embodiments of the present application provide an information recommendation method and device for a recommendation system, so as to coordinate between a recall module and a sorting module, so that the information in the recommendation list meets the information needs of the user and improves the user experience.

[0022] In order to enable those skilled in the art to better understand the technical solutions in the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0023] In some scenarios, the recommendation system recommends information that users are interested in to users based on their information needs and interests, thereby guiding users to discover their own information needs. When a user comes to the recommendation page, the system retrieves the information that the user is most likely to like from a large amount of candidate information based on the user's user data, and then goes through the recall module, sorting module, etc. to pull back several pieces of information that the user is most likely to like, forming a final recommendation list for display to the user.

[0024] For the recall module, its main function is to narrow the calculation scope of information, select the products that users are interested in from millions of pieces of information, and reduce the millions to hundreds or even dozens through simple models and algorithms. In this way, users will have the opportunity to get rapid information feedback with a delay of milliseconds.

[0025] For the sorting module, its main function is to obtain accurate sorting results, accurately sort the hundreds of information recalled from the recall module, and select the higher-ranked information to form a recommendation list for display to users.

[0026] For the recall module and the sorting module, there is collaboration between the multiple channels within the modules to ultimately determine the goals of the recall module and the sorting module. However, there is a lack of collaboration between the recall module and the sorting module, resulting in the final recommendation list being unrepresentative, the information recommended to users cannot meet user needs, and the user experience is low.

[0027] In order to solve the above technical problems, the present application provides an information recommendation method, device and electronic device of a recommendation system. The following is a detailed description of an information recommendation method, device and electronic device of a recommendation system provided by the present application in conjunction with the accompanying drawings.

[0028] like Figure 1 As shown, the embodiment of the present application provides an information recommendation method of a recommendation system, and the execution subject of the method can be a server, wherein the server can be an independent server or a server cluster composed of multiple servers. The information recommendation method of the recommendation system can specifically include the following steps S101-S104:

[0029] In S101, target data of a target user is obtained.

[0030] Specifically, the target user is a user determined by the server for information recommendation by the recommendation system, and the target data of the target user includes but is not limited to the user's own attribute data, user behavior data, user relationship data and user content data.

[0031] User-owned attribute data includes, but is not limited to, age, gender, occupation, address, education level, consumption level, etc. In this embodiment of the present application, the user-owned attribute data set is recorded as FUNC_INFO, and the user-owned attribute model can be recorded as FUNC_INFO (user_info1, user_info2, ..., user_infox), where x is the total number of sample data in the user-owned attribute data set. For example, user_info1 can represent age, user_info2 can represent gender, user_info3 can represent occupation, etc.

[0032] User behavior data includes but is not limited to: implicit feedback data and explicit feedback data. Implicit feedback data includes but is not limited to browsing, clicking, collecting, purchasing and other data. Explicit feedback data includes but is not limited to ratings and comments on items. Compared with explicit feedback data, implicit feedback data is more valuable, more usable, and has a relatively large amount of data. Therefore, in practical applications, implicit feedback data can be mainly used for analysis. The user behavior data set can be recorded as FUNC_BEHV, and the user behavior data model can be recorded as FUNC_BEHV (user_bhv1, user_bhv2, ..., user_bhvy), where y is the total number of sample data in the user behavior data set. For example, user_bhv1 can represent browsing, user_bhv2 can represent clicking, user_bhv3 can represent collecting, etc.

[0033] User relationship data includes but is not limited to: strong relationship data and weak relationship data. Strong relationship data includes mutual attention, mutual messaging, etc., and weak relationship data includes mutual likes, mutual comments, etc. Then the user relationship data set can be recorded as FUNC_RELT, and the user relationship data model can be recorded as FUNC_RELT(user_relt1, user_relt2, ..., user_reltz), where z is the total number of sample data in the user relationship data set. For example, user_relt1 can represent mutual attention, user_relt2 can represent mutual messaging, user_relt3 can represent mutual likes, etc.

[0034] User content data refers to content related to the user, including but not limited to other information related to the target user's current browsing information, other users' likes, comments, and attention to the target user, and other data.

[0035] The user content data set can be recorded as FUNC_CONT, and the user content data model can be recorded as FUNC_CONT (user_cont1, user_cont2, ..., user_contw), where w is the total number of sample data in the user content data set.

[0036] In S102, the user level to which the target user belongs is determined according to the target data.

[0037] Specifically, according to the above-mentioned user attribute data, user behavior data, user relationship data and user content data, the user's attributes and habits can be determined. According to the different attributes and habits of users, the user level can be divided into 4 levels, namely, the new user level, the growth level, the stable level and the decline level.

[0038] For new users, the goal is to increase retention. For growing users, the goal is to achieve rapid growth. For stable users, the goal is to achieve maturity and stability. For declining users, the goal is to prevent churn.

[0039] In S103, according to the user level, the objectives of the recall module and the ranking module included in the target object in the recommendation system are adjusted to the same collaborative objective.

[0040] Specifically, the collaborative goals include but are not limited to diversity and accuracy. Diversity refers to the richness of the information recommended to the user, that is, the variety of information. Accuracy means that the information recommended to the user highly meets the user's needs, that is, the recommended information is highly consistent with the information required by the user. Of course, according to actual conditions, the collaborative goals can also be other types, and the embodiments of this application are not limited here.

[0041] In the embodiment of the present application, the collaborative goals of the recall module and the sorting module are adjusted to be the same, so as to ensure the collaboration between the recall module and the sorting module. For example, for the new user level and the declining level, the collaborative goal of the recall module and the sorting module in the recommendation system is diversity. For growing and stable users, the collaborative goal of the recall module and the sorting module in the recommendation system is accuracy.

[0042] Combine the following Figure 2 The collaborative goals of the recall module and the sorting module of the recommendation system provided in the embodiment of the present application are explained.

[0043] like Figure 2 As shown, after determining the user's own attribute data, user behavior data, user relationship data and user content data, user stratification is performed. After user stratification, new household level, growth level, stable level and decline level are obtained. For the new household level, its goal is to improve retention, for the growth level, its goal is rapid growth, for the stable level, its goal is maturity and stability, and for the decline level, its goal is to prevent loss. For different user levels, the collaborative goals of the recall module and the sorting module are determined. For users at the new household level and the decline level, the collaborative goals of the recall module and the sorting module are diversity. For users at the growth level and the stable level, the collaborative goals of the recall module and the sorting module are accuracy.

[0044] In S104, target information is determined using the collaborative target, and the target information is recommended to the target user.

[0045] Specifically, after determining the collaborative goal, the recall module and the collaborative module are used to filter out information that meets the collaborative goal from millions of massive data as target information. A recommendation list is generated and recommended to the target user.

[0046] It can be seen from the technical solution provided in the above embodiments of the present application that the collaborative goals of the recall module and the sorting module are determined to be consistent, so as to achieve the purpose of collaboration between the recall module and the sorting module. The target information in the recommendation list is also more likely to meet the user's information needs, thereby improving the user experience.

[0047] In one possible implementation, Figure 3 As shown, the specific processing method of the above S102 can be various, and an optional implementation method is provided below. For details, please refer to the processing of the following S1020-S1023.

[0048] In S1020, the target data is converted into a numerical feature vector.

[0049] Specifically, logistic regression is used to map user attribute data, user behavior data, user relationship data, and user content data to a value between 0 and 1 to obtain a numerical feature vector.

[0050] In S1021 , a first weight parameter corresponding to a numerical feature vector of each type of target data and a second weight parameter of a single sample data in each type of target data are determined.

[0051] Specifically, there are four types of target data: user attribute data, user behavior data, user relationship data, and user content data. For each type of data, a first weight parameter is fitted for each type of data according to its contribution to the collaborative goal through logistic regression. For example, among the above four types of data, user behavior data and user relationship data are the main data for determining the collaborative goal, followed by user content data, and user attribute data is the last. Therefore, the first weight parameters of user behavior data and user relationship data can be set to be the same or similar, the first weight parameters of user behavior data and user relationship data are higher than the first weight parameter of user content data, and the first weight parameter of user content data is higher than the first weight parameter of user attribute data.

[0052] Furthermore, for each type of data, the contribution of the sample data to the collaborative goal is also different. For example, user behavior data includes browsing, clicking, collecting, purchasing, rating, and evaluation data. Among them, browsing, clicking, collecting, and purchasing data as implicit feedback data are more representative than rating and evaluation data as explicit feedback data. Therefore, the second weight parameter of browsing, clicking, collecting, and purchasing data can be set higher than the second weight parameter of rating and evaluation.

[0053] In the embodiment of the present application, the first weight parameter of the user's own attribute data is recorded as beta_info, the first weight parameter of the user's behavior data is recorded as beta_bhv, the first weight parameter of the user's relationship data is recorded as beta_relt, and the first weight parameter of the user's content data is recorded as beta_cont. The second weight parameter for a single sample data in the user's own attribute data set is recorded as alph_infox, where x is the number of sample data in the user's own attribute data set. The second weight parameter for a single sample data in the user's behavior data set is recorded as alph_bhvy, where y is the number of sample data in the user's behavior data set. The second weight parameter for a single sample data in the user's relationship data set is recorded as alph_reltz, where z is the number of sample data in the user's relationship data set. The second weight parameter for a single sample data in the user's content data set is recorded as alph_contw, where w is the number of sample data in the user's content data set.

[0054] In S1022 , a first score is calculated based on the first weight parameter, the second weight parameter, and the numerical feature vector.

[0055] In one possible implementation, calculating the first score includes: for each type of target data, respectively calculating the first product of the numerical feature vector of each sample data in each type of target data and the second weight parameter corresponding to each sample data; respectively superimposing the first products in each type of target data to obtain the first sum of each type of target data; multiplying the first sum of each type of target data by the first weight parameter corresponding to each type of target data to obtain the second product of each type of target data; superimposing the second products of each type of target data to obtain a second sum, and the second sum is the first score.

[0056] Specifically, firstly, the first product of the numerical feature vector of each sample data of each type of data and the second weight parameter corresponding thereto is obtained, and then the first products of each type of data are superimposed to obtain a first sum value. The first sum value of each type of data is multiplied by the first weight parameter corresponding to each type of data to obtain a second product, and the second products of each type of data are superimposed to obtain a second sum value, which is the first score.

[0057] In the embodiment of the present application, the first score is recorded as USER_LEVEL_SCORE, and USER_LEVEL_SCORE can be specifically calculated using the following formula:

[0058] USER_LEVEL_SCORE=beta_info*(alph_info1*user_info1+alph_info2*user_info2+…+alph_infox*user_infox)+

[0059] beta_bhv*(alph_bhv1*user_bhv1+alph_bhv2*user_bhv2+…+alph_bhvy*user_bhvy)+

[0060] beta_relt*(alph_relt1*user_relt1+alph_relt2*user_relt2+…+alph_reltz*user_re ltz)+

[0061] beta_cont*(alph_cont1*user_cont1+alph_cont2*user_cont2+…+alph_contw*user_contw)

[0062] Among them, "*" represents multiplication, alph_info1 represents the second weight parameter of the first sample data (such as gender) in the user's own attribute data, and user_info1 is the first sample data in the user's own attribute data or the numerical feature vector of the first sample data. alph_info2 represents the second weight parameter of the second sample data (such as age) in the user's own attribute data, and user_info2 is the second sample data in the user's own attribute data or the numerical feature vector of the second sample data. alph_infox represents the second weight parameter of the x-th sample data in the user's own attribute data, and user_infox is the x-th sample data in the user's own attribute data or the numerical feature vector of the x-th sample data.

[0063] alph_bhv1 represents the second weight parameter of the first sample data (such as click) in the user behavior data, and user_bhv1 is the first sample data in the user behavior data or the numerical feature vector of the first sample data. alph_bhv2 represents the second weight parameter of the second sample data (such as browse) in the user behavior data, and user_bhv2 is the second sample data in the user behavior data or the numerical feature vector of the second sample data. alph_bhvy represents the second weight parameter of the yth sample data in the user behavior data, and user_bhvy is the yth sample data in the user behavior data or the numerical feature vector of the yth sample data.

[0064] alph_relt1 represents the second weight parameter of the first sample data in the user relationship data (such as mutual attention), and user_relt1 is the first sample data in the user relationship data or the numerical feature vector of the first sample data. alph_relt2 represents the second weight parameter of the second sample data in the user relationship data (such as mutual messaging), and alph_relt2 is the second sample data in the user relationship data or the numerical feature vector of the second sample data. alph_reltz represents the second weight parameter of the zth sample data in the user relationship data, and alph_reltz is the zth sample data in the user relationship data or the numerical feature vector of the zth sample data.

[0065] alph_cont1 represents the second weight parameter of the first sample data in the user content data, user_cont1 is the first sample data in the user content data or the numerical feature vector of the first sample data. alph_cont2 represents the second weight parameter of the second sample data in the user content data, user_cont2 is the second sample data in the user content data or the numerical feature vector of the second sample data. alph_contw represents the second weight parameter of the wth sample data in the user content data, user_contw is the wth sample data in the user content data or the numerical feature vector of the wth sample data.

[0066] In S1023, the user level to which the target user belongs is determined according to the level range interval in which the first score is located, and the level range interval corresponds to the user level.

[0067] Among them, corresponding to the first score calculated by the above formula, the corresponding user level is determined according to the level range interval in which it is located. For example, the first level range interval corresponds to the new user level, the second level range interval corresponds to the growth level, the third level range interval corresponds to the stable level, and the fourth level range interval corresponds to the declining level.

[0068] Through the technical solution disclosed in the embodiments of the present application, the score value of the target user is calculated through the user's target data, and the corresponding user level is determined according to the range of the score value. Targeted recommendation information is provided for users of different levels, which can accurately meet user needs and improve user experience.

[0069] In one possible implementation, Figure 4 As shown, the specific processing method of the above S103 can be various, and an optional implementation method is provided below. For details, please refer to the processing of the following S1030-S1031.

[0070] In S1030 , a second score of the target object is determined according to the user level.

[0071] In one possible implementation, S1030 includes: determining, according to the user level of the target user, a third weight parameter between the target object and the preset first collaborative target, and a fourth weight parameter between the target object and the preset second collaborative target; determining, according to the user level of the target user, the collaborative parameter corresponding to the target object; and determining the second score of the target object based on the third weight parameter, the fourth weight parameter and the collaborative parameter.

[0072] Specifically, the first collaborative target may be diversity, and the second collaborative target may be accuracy. The target objects may be a recall module and a sorting module. The third weight parameter of the first collaborative target of the recall module may be recorded as RECALL_DIVERSITY_user_level, and the fourth weight parameter of the second collaborative target of the recall module may be recorded as RECALL_PRECISE_user_level. The third weight parameter of the first collaborative target of the sorting module may be recorded as RANK_DIVERSITY_user_level, and the fourth weight parameter of the second collaborative target of the sorting module may be recorded as RANK_PRECISE_user_level. The collaborative parameters may be the parameter PARA_A_user_level of the first collaborative target of the recall module, the parameter PARA_B_user_level of the second collaborative target of the recall module, the parameter PARA_C_user_level of the first collaborative target of the sorting module, and the parameter PARA_D_user_level of the second collaborative target of the sorting module. In a possible implementation, the collaborative parameter is consistent with the first score USER_LEVEL_SCORE calculated in the above embodiment when the user is stratified.

[0073] In one possible implementation, calculating the second score includes: calculating a first product of the collaborative parameters of the first collaborative target of the recall module and the third weight parameter of the first collaborative target of the recall module; calculating a second product of the collaborative parameters of the second collaborative target of the recall module and the fourth weight parameter of the second collaborative target of the recall module; calculating a third product of the collaborative parameters of the first collaborative target of the sorting module and the third weight parameter of the first collaborative target of the sorting module; calculating a fourth product of the collaborative parameters of the second collaborative target of the sorting module and the fourth weight parameter of the second collaborative target of the sorting module; superimposing the first product, the second product, the third product and the fourth product to obtain a third sum, and the third sum is the second score.

[0074] Furthermore, the second score value (RECALL_RANK_COOPERATE) can be calculated according to the following formula:

[0075] RECALL_RANK_COOPERATE=

[0076] PARA_A_user_level*RECALL_DIVERSITY_user_level+

[0077] PARA_B_user_level*RECALL_PRECISE_user_level+

[0078] PARA_C_user_level*RANK_DIVERSITY_user_level+

[0079] PARA_D_user_level*RANK_PRECISE_user_level.

[0080] In S1031, based on the target range interval where the second score is located, a collaborative target is determined, the target range interval corresponds to the collaborative target, and the targets of the recall module and the sorting module included in the target object are adjusted to be the collaborative target.

[0081] In one possible implementation, S1031 includes: when the second score is within the first target range, determining the first collaborative target corresponding to the first target range as the collaborative target; when the second score is within the second target range, determining the second collaborative target corresponding to the second target range as the collaborative target.

[0082] Specifically, the second scores of the recall module and the sorting module are calculated according to the target user layers, and the emphasis of the recall module and the sorting module is determined according to the second scores. That is, the collaborative target is determined according to the range of the second scores, so that the emphasis of the recall module and the sorting module corresponding to users at different levels is different.

[0083] Through the technical solution disclosed in the embodiment of the present application, by calculating the second scores for users at different levels and determining the focus of the recall module and the sorting module according to the range of the second scores, the recommended information can meet user needs and improve the user experience.

[0084] Corresponding to the information recommendation method of the recommendation system provided in the above embodiment, based on the same technical concept, the embodiment of the present application also provides an information recommendation device of the recommendation system, Figure 5 This is a schematic diagram of the module composition of the information recommendation device of the recommendation system provided in the embodiment of the present application. The information recommendation device of the recommendation system is used to perform Figure 1 and Figure 4 The information recommendation method of the recommendation system described in Figure 5As shown, the information recommendation device 5 of the recommendation system includes: an acquisition module 501 , a first determination module 502 , a second determination module 503 and a recommendation module 504 .

[0085] An acquisition module 501 is used to acquire target data of a target user; a first determination module 502 is used to determine the user level to which the target user belongs based on the target data; a second determination module 503 is used to adjust the goals of a recall module and a sorting module included in the target object in the recommendation system to the same collaborative goal based on the user level; a recommendation module 504 is used to determine target information using the collaborative goal and recommend the target information to the target user.

[0086] Through a technical solution disclosed in an embodiment of the present application, the collaborative targets of the recall module and the sorting module are determined to be consistent, thereby achieving the purpose of collaboration between the recall module and the sorting module. The target information in the recommendation list is also more likely to meet the user's information needs, thereby improving the user experience.

[0087] In a possible implementation, the first determination module 502 is also used to convert the target data into a numerical feature vector; determine the first weight parameter corresponding to the numerical feature vector of each type of target data and the second weight parameter of the single sample data in each type of target data according to the numerical feature vector; calculate the first score based on the first weight parameter, the second weight parameter and the numerical feature vector; determine the user level to which the target user belongs according to the level range interval in which the first score is located, and the level range interval corresponds to the user level.

[0088] In one possible implementation, the first determination module 502 is further used to calculate, for each type of target data, the first product of the numerical feature vector of each sample data in each type of target data and the second weight parameter corresponding to each sample data; superimpose the first products in each type of target data to obtain the first sum of each type of target data; multiply the first sum of each type of target data by the first weight parameter corresponding to each type of target data to obtain the second product of each type of target data; superimpose the second products of each type of target data to obtain a second sum, which is the first score.

[0089] In one possible implementation, the second determination module 503 is also used to determine the second score of the target object according to the user level; determine the collaborative target based on the target range interval in which the second score is located, and the target range interval corresponds to the collaborative target; and adjust the targets of the recall module and the sorting module included in the target object to be collaborative targets.

[0090] In one possible implementation, the second determination module 503 is further used to determine, according to the user level of the target user, a third weight parameter between the target object and the preset first collaborative target, and a fourth weight parameter between the target object and the preset second collaborative target; determine the collaborative parameter corresponding to the target object according to the user level of the target user; and determine the second score of the target object based on the third weight parameter, the fourth weight parameter and the collaborative parameter.

[0091] In one possible implementation, the second determination module 503 is further used to determine, when the second score is within the first target range, that the first collaborative target corresponding to the first target range is the collaborative target; and when the second score is within the second target range, determine, when the second score is within the second target range, that the second collaborative target corresponding to the second target range is the collaborative target.

[0092] The second determination module 503 is also used to calculate the first product of the collaborative parameter of the first collaborative target of the recall module and the third weight parameter of the first collaborative target of the recall module; calculate the second product of the collaborative parameter of the second collaborative target of the recall module and the fourth weight parameter of the second collaborative target of the recall module; calculate the third product of the collaborative parameter of the first collaborative target of the sorting module and the third weight parameter of the first collaborative target of the sorting module; calculate the fourth product of the collaborative parameter of the second collaborative target of the sorting module and the fourth weight parameter of the second collaborative target of the sorting module; superimpose the first product, the second product, the third product and the fourth product to obtain a third sum, and the third sum is the second score.

[0093] The information recommendation device of the recommendation system provided in the embodiment of the present application can implement each process in the embodiment corresponding to the information recommendation method of the recommendation system mentioned above, and will not be described again here to avoid repetition.

[0094] It should be noted that the information recommendation device of the recommendation system provided in the embodiment of the present application and the information recommendation method of the recommendation system provided in the embodiment of the present application are based on the same inventive concept. Therefore, the specific implementation of this embodiment can refer to the implementation of the information recommendation method of the aforementioned recommendation system, and the repeated parts will not be repeated.

[0095] Corresponding to the information recommendation method of the recommendation system provided in the above embodiment, based on the same technical concept, the embodiment of the present application further provides an electronic device, which is used to execute the information recommendation method of the recommendation system. Figure 6 A schematic diagram of the structure of an electronic device for implementing various embodiments of the present invention is shown in FIG. Figure 6As shown. The electronic device may have relatively large differences due to different configurations or performances, and may include one or more processors 601 and memory 602, and the memory 602 may store one or more storage applications or data. Among them, the memory 602 can be a temporary storage or a permanent storage. The application stored in the memory 602 may include one or more modules (not shown in the figure), and each module may include a series of computer executable instructions in the electronic device. Furthermore, the processor 601 can be configured to communicate with the memory 602 to execute a series of computer executable instructions in the memory 602 on the electronic device. The electronic device may also include one or more power supplies 603, one or more wired or wireless network interfaces 604, one or more input and output interfaces 605, and one or more keyboards 606.

[0096] In this embodiment, the electronic device includes a processor, a communication interface, a memory and a communication bus; wherein the processor, the communication interface and the memory communicate with each other through the bus; the memory is used to store computer programs; the processor is used to execute the programs stored in the memory to implement the following method steps:

[0097] Obtain target data of the target user; determine the user level to which the target user belongs based on the target data; adjust the goals of the recall module and the sorting module included in the target object in the recommendation system to the same collaborative goal based on the user level; determine the target information using the collaborative goal, and recommend the target information to the target user.

[0098] Through a technical solution provided by the present application, the collaborative goals of the recall module and the sorting module are determined to be consistent, so as to achieve the purpose of collaboration between the recall module and the sorting module. The target information in the recommendation list is also more likely to meet the user's information needs, thereby improving the user experience.

[0099] In a specific embodiment, the embodiment of the present application provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented:

[0100] Obtain target data of the target user; determine the user level to which the target user belongs based on the target data; adjust the goals of the recall module and the sorting module included in the target object in the recommendation system to the same collaborative goal based on the user level; determine the target information using the collaborative goal, and recommend the target information to the target user.

[0101] Through a technical solution provided by the present application, the collaborative goals of the recall module and the sorting module are determined to be consistent, so as to achieve the purpose of collaboration between the recall module and the sorting module. The target information in the recommendation list is also more likely to meet the user's information needs, thereby improving the user experience.

[0102] In a specific embodiment, the embodiment of the present application provides a chip, the chip includes a processor and a communication interface, the communication interface is coupled to the processor, and the processor is used to run a program or instruction to implement the following steps:

[0103] Obtain target data of the target user; determine the user level to which the target user belongs based on the target data; adjust the goals of the recall module and the sorting module included in the target object in the recommendation system to the same collaborative goal based on the user level; determine the target information using the collaborative goal, and recommend the target information to the target user.

[0104] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, devices, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0105] The present invention is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of the processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0106] These computer program instructions may also be stored in a computer readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0107] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0108] In a typical configuration, an electronic device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0109] The memory may include non-permanent storage in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0110] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0111] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0112] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, devices or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment or an embodiment combining software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

[0113] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included within the scope of the claims of the present application.

Claims

1. An information recommendation method for a recommendation system, characterized in that: The method comprises: Obtain target data of target users; Determining, according to the target data, the user level to which the target user belongs; According to the user level, adjusting the objectives of the recall module and the ranking module included in the target object in the recommendation system to the same collaborative objective; Determine target information using the collaborative target, and recommend the target information to the target user; The step of adjusting the objectives of the recall module and the sorting module included in the target object of the recommendation system to the same collaborative objective according to the user level includes: Determining a second score of the target object according to the user level; Determining the collaborative target based on the target range in which the second score is located, the target range corresponding to the collaborative target; The targets of the recall module and the sorting module included in adjusting the target object are both the collaborative targets.

2. The method according to claim 1, characterized in that The target data includes at least two different types of data, the types of data include: user attribute data, user behavior data, user relationship data and user content data; Determining the user level to which the target user belongs according to the target data includes: Converting the target data into a numerical feature vector; Determine a first weight parameter corresponding to a numerical feature vector of each type of target data, and a second weight parameter of a single sample data in each type of target data; Calculating a first score based on the first weight parameter, the second weight parameter, and the numerical feature vector; The user level to which the target user belongs is determined according to the level range interval in which the first score is located, and the level range interval corresponds to the user level.

3. The method according to claim 2, characterized in that The calculating a first score based on the first weight parameter, the second weight parameter and the numerical feature vector comprises: For each type of target data, respectively calculate the first product of the numerical feature vector of each sample data in the current type of target data and the second weight parameter corresponding to each sample data; and superimpose the first products in the current type of target data to obtain the first sum value of the current type of target data; Multiplying the first sum value of each type of target data and the first weight parameter corresponding to each type of target data to obtain a second product of each type of target data; The second products of the target data of each type are superimposed to obtain a second sum value, where the second sum value is the first score.

4. The method according to claim 1, characterized in that: Determining the second score of the target object according to the user level includes: According to the user level of the target user, respectively determine a third weight parameter of the target object in the preset first collaborative target and a fourth weight parameter of the target object in the preset second collaborative target; Determining a collaboration parameter corresponding to the target object according to the user level of the target user; determining a second score of the target object according to the third weight parameter, the fourth weight parameter and the coordination parameter; The determining the collaborative target based on the target range of the second score includes: When the second score is within a first target range, determining a first collaborative target corresponding to the first target range as the collaborative target; When the second score is within a second target range, a second collaborative target corresponding to the second target range is determined as the collaborative target.

5. The method according to claim 4, characterized in that Determining the second score of the target object according to the third weight parameter, the fourth weight parameter and the collaborative parameter comprises: Calculating a first product of a coordination parameter corresponding to the recall module and a third weight parameter of the recall module at the first coordination target; Calculating a second product of a coordination parameter corresponding to the recall module and a fourth weight parameter of the recall module at the second coordination target; Calculating a third product of a coordination parameter corresponding to the sorting module and a third weight parameter of the sorting module at the first coordination target; Calculating a fourth product of a coordination parameter corresponding to the sorting module and a fourth weight parameter of the sorting module at the second coordination objective; The first product, the second product, the third product and the fourth product are superimposed to obtain a third sum value, and the third sum value is the second score.

6. An information recommendation device of a recommendation system, characterized in that: The device comprises: An acquisition module is used to acquire target data of target users; A first determination module, configured to determine the user level to which the target user belongs according to the target data; A second determination module is used to adjust the objectives of a recall module and a ranking module included in the target object in the recommendation system to the same collaborative objective according to the user level; A recommendation module, used to determine target information using the collaborative target, and recommend the target information to the target user; The second determination module is specifically used to determine the second score of the target object according to the user level; determine the collaborative target based on the target range interval in which the second score is located, and the target range interval corresponds to the collaborative target; adjust the targets of the recall module and the sorting module included in the target object to be the collaborative target.

7. An electronic device comprising a processor, a communication interface, a memory and a communication bus; wherein: The processor, the communication interface and the memory communicate with each other via a bus; the memory is used to store computer programs; the processor is used to execute the programs stored in the memory to implement the information recommendation method steps of the recommendation system as described in any one of claims 1 to 5.

8. A computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the steps of the information recommendation method of the recommendation system according to any one of claims 1 to 5 are implemented.

Images