Information recommendation method and device, electronic device, and storage medium

Abstract

The embodiments of the present application disclose an information recommendation method, an information recommendation device, an electronic device, and a storage medium. The method comprises: obtaining multiple recommendation features, the multiple recommendation features including target object features, virtual object features corresponding to multiple virtual objects to be recommended, and attribute sequence pattern features associated with the attribute sequence of the target object, wherein the attribute sequence of the target object includes the attribute sequence of the target object for different virtual objects, the attribute sequence includes multiple attribute identifiers arranged in chronological order, and the attribute identifiers are used to characterize the historical operation attribute types of the target object for the corresponding virtual objects; fusing the multiple recommendation features to obtain a fused feature, and determining the recommendation probability of the virtual object to the target object based on the fused feature; and recommending to the target object virtual objects whose recommendation probability is greater than a preset threshold. The technical solution of the embodiments of the present application can improve the accuracy of information recommendation.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and more specifically, to an information recommendation method and device, an electronic device, and a computer-readable storage medium. Background Art

[0002] With the continuous development of artificial intelligence (AI), its application in intelligent recommendations is becoming increasingly widespread. For example, AI is being used to provide personalized product recommendations and targeted advertising for target audiences. The key to achieving these personalized product recommendations and targeted advertising lies in more precise understanding of the target audience's individual needs. The more accurately we understand the target audience's needs, the more accurate the recommended items or information will be. Summary of the Invention

[0003] To solve the above technical problems, embodiments of the present application provide an information recommendation method and information recommendation device, an electronic device, and a storage medium, which can improve the accuracy of information recommendation.

[0004] Other features and advantages of the present application will become apparent from the following detailed description, or may be learned in part by practice of the present application.

[0005] According to one aspect of an embodiment of the present application, an information recommendation method is provided, comprising: obtaining a plurality of recommendation features, the plurality of recommendation features including target object features, virtual object features corresponding to a plurality of virtual objects to be recommended, and attribute sequence pattern features associated with the attribute sequence of the target object, wherein the attribute sequence of the target object includes the attribute sequences of the target object for different virtual objects respectively, the attribute sequence includes a plurality of attribute identifiers arranged in chronological order, the attribute identifiers being used to characterize the historical operation attribute types of the target object for the corresponding virtual objects; fusing the plurality of recommendation features to obtain a fused feature, and determining a recommendation probability of the virtual object to the target object based on the fused feature; and recommending to the target object a virtual object whose recommendation probability is greater than a preset threshold.

[0006] According to one aspect of an embodiment of the present application, an information recommendation device is provided, including: a feature acquisition module for multiple recommendation features, the multiple recommendation features including target object features, virtual object features corresponding to multiple virtual objects to be recommended, and attribute sequence pattern features associated with the attribute sequence of the target object, wherein the attribute sequence of the target object includes the attribute sequences of the target object for different virtual objects, the attribute sequence includes multiple attribute identifiers arranged in chronological order, and the attribute identifiers are used to characterize the historical operation attribute types of the target object for the corresponding virtual objects; a probability acquisition module, connected to the feature acquisition module, for fusing the multiple recommendation features to obtain a fused feature, and determining the recommendation probability of the virtual object recommended to the target object based on the fused feature; a recommendation module for recommending to the target object a virtual object whose recommendation probability is greater than a preset threshold.

[0007] According to one aspect of an embodiment of the present application, an electronic device is provided, including a processor and a memory, wherein computer-readable instructions are stored in the memory, and when the computer-readable instructions are executed by the processor, the above information recommendation method is implemented.

[0008] According to one aspect of an embodiment of the present application, a computer-readable storage medium is provided, on which computer-readable instructions are stored. When the computer-readable instructions are executed by a processor of a computer, the computer executes the above information recommendation method.

[0009] According to one aspect of an embodiment of the present application, a computer program product or computer program is provided, comprising computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the information recommendation methods provided in the various optional embodiments described above.

[0010] In the technical solution provided in the embodiments of the present application, the target object features, the virtual object features corresponding to the multiple virtual objects to be recommended, and the attribute sequence pattern features associated with the target object attribute sequence are fused, wherein the target object attribute sequence includes the target object's attribute sequences for different virtual objects, and the recommendation probability of recommending the virtual object to the target object is determined based on the obtained fused features. Because the attribute sequences of the same target object for different virtual objects are used as an important factor in determining the recommendation probability of recommending the virtual object to the target object, it is possible to more accurately explore the target object's deeper preferences for the multiple virtual objects to be recommended, improve the accuracy of the obtained recommendation probability, and thus improve the accuracy of the information recommendation.

[0011] It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] The accompanying drawings are incorporated into and constitute a part of the specification, illustrating embodiments consistent with the present application and, together with the specification, serving to explain the principles of the present application. It is obvious that the drawings described below are merely some embodiments of the present application, and a person of ordinary skill in the art can derive other drawings based on these drawings without inventive effort. In the drawings:

[0013] Figure 1 is a flowchart of the information recommendation method shown in an exemplary embodiment of the present application;

[0014] Figure 2 yes Figure 1 A flowchart of an exemplary embodiment of step S100 in the illustrated embodiment;

[0015] Figure 3 yes Figure 2 A flowchart of an exemplary embodiment of step S120 in the illustrated embodiment;

[0016] Figure 4 yes Figure 3 A flowchart of an exemplary embodiment of step S121 in the embodiment shown;

[0017] Figure 5 yes Figure 3 A flowchart of an exemplary embodiment of step S122 in the embodiment shown;

[0018] Figure 6 yes Figure 1 A flowchart of an exemplary embodiment of step S200 in the illustrated embodiment;

[0019] Figure 7 is a block diagram of an information recommendation device shown in an exemplary embodiment of the present application;

[0020] Figure 8 A schematic diagram of the structure of a computer system suitable for implementing an electronic device according to an embodiment of the present application is shown. DETAILED DESCRIPTION

[0021] Exemplary embodiments will be described in detail herein, with examples illustrated in the accompanying drawings. When the following description refers to the drawings, identical numerals in different figures represent identical or similar elements unless otherwise indicated. The embodiments described in the following exemplary embodiments are not intended to represent all embodiments consistent with the present application. Rather, they are merely examples of apparatus and methods consistent with certain aspects of the present application, as detailed in the appended claims.

[0022] The block diagrams shown in the accompanying drawings are merely functional entities and do not necessarily correspond to physically separate entities. That is, these functional entities may be implemented in software, in one or more hardware modules or integrated circuits, or in different networks and / or processor devices and / or microcontroller devices.

[0023] The flowcharts shown in the accompanying drawings are for illustrative purposes only and do not necessarily include all contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps may be decomposed, while others may be combined or partially combined. Therefore, the actual execution order may vary depending on the actual situation.

[0024] It should also be noted that the term "plurality" used in this application refers to two or more. "And / or" describes the relationship between associated objects, indicating that three possible relationships exist. For example, "A and / or B" can mean: A exists alone, A and B exist simultaneously, or B exists alone. The character " / " generally indicates that the associated objects are in an "or" relationship.

[0025] Artificial Intelligence (AI) refers to the theories, methods, techniques, and application systems that use digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, to perceive the environment, acquire knowledge, and use that knowledge to achieve optimal results. In other words, AI is a comprehensive technology within computer science that seeks to understand the essence of intelligence and produce new intelligent machines that can respond in a manner similar to human intelligence. AI also involves studying the design principles and implementation methods of various intelligent machines, enabling them to possess the capabilities of perception, reasoning, and decision-making.

[0026] Artificial intelligence (AI) technology is a comprehensive discipline encompassing a wide range of fields, encompassing both hardware and software technologies. Foundational AI technologies generally include sensors, specialized AI chips, cloud computing, distributed storage, big data processing, operating / interaction systems, and mechatronics. AI software technologies primarily encompass computer vision, speech processing, natural language processing, and machine learning / deep learning.

[0027] Machine learning (ML) is a multidisciplinary field that encompasses probability theory, statistics, approximation theory, convex analysis, and algorithmic complexity theory. It specifically studies how computers can simulate or implement human learning behaviors to acquire new knowledge or skills and reorganize existing knowledge structures to continuously improve their performance. Machine learning is at the core of artificial intelligence and the fundamental way to make computers intelligent. Its applications span all areas of AI. Machine learning and deep learning typically include techniques such as artificial neural networks, belief networks, reinforcement learning, transfer learning, inductive learning, and learning through demonstration.

[0028] The information recommendation method and device, electronic device, and computer-readable storage medium proposed in the embodiments of the present application involve artificial intelligence technology and machine learning technology, and these embodiments will be described in detail below.

[0029] See also Figure 1 , Figure 1 This is a flowchart of the information recommendation method shown in an exemplary embodiment of the present application. Figure 1 As shown, an information recommendation method involved in this application includes:

[0030] Step S100: Acquire multiple recommendation features.

[0031] The multiple recommendation features include target object features, virtual object features corresponding to multiple virtual objects to be recommended, and attribute sequence pattern features associated with the attribute sequence of the target object.

[0032] In this embodiment, a virtual object refers to a virtual resource to be recommended to a target object, such as a virtual commodity or virtual content on an online e-commerce platform, such as an advertisement. A target object exists in the form of an object identifier, for example, a user account on an online e-commerce platform is used to identify the corresponding target object.

[0033] Target object characteristics include basic attribute characteristics of the target object, such as age, gender, education level, city level, etc.; target object consumption characteristics, such as the total number of payments, total amount, distribution of payment numbers within a certain time period (within 24 hours, within a week, within a month, within half a year), distribution of payment amounts, average amount per payment, etc.; target object attribute characteristics such as browsing time, number of page clicks, etc.

[0034] Virtual object characteristics include basic attribute characteristics of virtual objects such as category, price, brand, purchase rating, comment sentiment, etc., and consumption characteristics of virtual objects such as the number of times an item has been purchased, the number of times it has been clicked and viewed, the number of times it has been added to the shopping cart, and the number of times similar items have been purchased.

[0035] Optionally, the target object features and / or the virtual object features are preprocessed.

[0036] The preprocessing process may include at least one of missing value processing and outlier processing.

[0037] Missing value handling can include:

[0038] Missing value filling: continuous features are filled with the mean, and discrete features are filled with constants as separate categories.

[0039] Discard features with too many missing values: Set the missing value filtering threshold = sample data volume * k, where k is the missing filtering rate, which is proportional to the sample data volume and is set according to the application scenario. If the number of missing data for a feature exceeds the missing value filtering threshold, filter this feature and delete the single-value feature at the same time.

[0040] Outlier processing: Based on the feature distribution, outliers with large feature values ​​and ranked in the top 1 / m are discarded. m can be set to 10000, depending on the application scenario.

[0041] In this embodiment, the attribute sequence of the target object includes the attribute sequences of the target object for different virtual objects. The attribute sequence includes multiple attribute identifiers arranged in chronological order. The attribute identifiers are used to represent the historical operation attribute types of the target object for the virtual objects.

[0042] The target object's historical operation attributes on the virtual object can directly reflect its preference for the virtual object. Therefore, this embodiment uses the attribute sequence pattern feature associated with the target object's attribute sequence as a recommendation feature to improve recommendation accuracy.

[0043] The target object's historical operation attributes on virtual objects are divided into direct operation attributes and indirect operation attributes. Direct operation attributes are the target object's direct operation attributes on virtual objects, such as the target object's direct click, purchase, favorite, and shopping cart attributes on a virtual commodity on an electronic trading platform. Indirect operation attributes are some operation attributes that are not directly related to the virtual commodity and are performed by the target object in order to achieve the purpose of directly operating the virtual commodity. For example, in order to purchase a virtual commodity on an electronic trading platform, the target object first registers and logs in to the electronic trading platform. Then, this series of operation attributes of the target object registering and logging in to the electronic trading platform are the target object's indirect operation attributes on the virtual commodity.

[0044] Step S200: A fusion feature is obtained by fusing multiple recommendation features, and a recommendation probability of recommending a virtual object to a target object is determined based on the fusion feature.

[0045] Feature fusion is the process of fusing multiple features together. The existing feature fusion algorithms can be roughly divided into the following three categories: one is simple feature combination, that is, all feature vectors are combined together in a serial or parallel manner to form a new feature vector; the second is feature selection, that is, from the newly combined feature vector, a data that is optimal for classification is selected from each corresponding dimension of data, and finally the selected data is combined into a new feature; the last one is feature transformation, that is, all feature vectors are put together, and then transformed into a new feature using certain mathematical methods.

[0046] It can be seen that the fusion feature contains relevant feature information of multiple recommendation features, that is, feature information related to the target object and feature information related to the virtual object. Based on this feature information, the recommendation probability of recommending the virtual object to the target object can be comprehensively measured to improve the accuracy of the recommendation probability.

[0047] Step S300: recommending a virtual object whose recommendation probability is greater than a preset threshold to the target object.

[0048] In the technical solution provided in the embodiments of the present application, target object features, virtual object features corresponding to multiple virtual objects to be recommended, and attribute sequence pattern features associated with the target object's attribute sequence are fused, wherein the target object's attribute sequence includes the target object's attribute sequences for different virtual objects. The recommendation probability of recommending the virtual object to the target object is determined based on the obtained fused features. Because the attribute sequences of the same target object for different virtual objects are used as an important factor in determining the recommendation probability of recommending the virtual object to the target object, it is possible to more accurately discover the target object's more accurate preferences for multiple virtual objects to be recommended, improve the accuracy of the obtained recommendation probability, and thereby improve the accuracy of the information recommendation.

[0049] Optional, see Figure 2 , Figure 2 yes Figure 1 In the embodiment shown, step S100 is a flowchart of an exemplary embodiment, as shown in FIG. Figure 2 As shown, step S100 includes the following steps:

[0050] Step S110: Obtain attribute sequences of the target object for different virtual objects.

[0051] In this embodiment, in order to mine the target object's preference patterns for multiple different virtual objects to be recommended, multiple attribute sequences of the target object for different virtual objects may be obtained in advance.

[0052] For example, target user Xiao Ming clicks button a on page A to enter page B, then browses for a while and clicks button b to enter page C. Target user Xiao Li clicks button a on page A to enter page B, then browses for a while and clicks button c to return to page A. The browsing sequence of target user Xiao Ming can be labeled as: AaBbC, and the browsing sequence of target user Xiao Li can be labeled as: AaBcA. Sequence information is sequential, so patterns can be mined using sequential pattern mining algorithms.

[0053] Step S120: performing sequence pattern mining on all acquired attribute sequences to obtain attribute sequence pattern features associated with the attributes of the target object.

[0054] Sequence pattern mining involves mining for patterns that occur frequently relative to time or other patterns. Given a set of distinct sequences, where each sequence consists of distinct elements arranged in an ordered sequence, and each element (transaction) consists of distinct items, and a target object with a specified minimum support threshold, sequence pattern mining aims to find all frequent subsequences within the set of sequences whose frequency of occurrence is at least the specified minimum support threshold.

[0055] After obtaining the attribute sequences of the same target object for different virtual objects, this embodiment performs sequence mining on all attribute sequences, mining attribute subsequences with an occurrence frequency greater than a certain threshold. These subsequences are used as attribute sequence pattern features that can reflect the target object's preferences for different virtual objects. Furthermore, using the fusion features obtained by fusing the target object features, the virtual object features corresponding to multiple virtual objects to be recommended, and the attribute sequence pattern features related to the target object, the probability of recommending the virtual object to the target object is calculated, and virtual objects with a recommendation probability greater than a preset threshold are recommended to the target object. The information recommendation method provided by this embodiment can mine effective information that can reflect the target object's attribute preferences from the target object's multiple attribute sequence data for multiple virtual objects, thereby improving the accuracy of information recommendation.

[0056] See Figure 3 , Figure 3 yes Figure 2 In the embodiment shown, step S120 is a flowchart of an exemplary embodiment, as shown in FIG. Figure 3 As shown, step S120 may include the following steps:

[0057] Step S121: traverse a prefix in all attribute sequences, take a prefix as a designated prefix, and obtain a projection data set corresponding to the designated prefix. The projection data set corresponding to the designated prefix includes the attribute identification string located after the corresponding prefix in all attribute sequences.

[0058] It's important to note that if an attribute sequence includes multiple identical prefixes, the attribute identifier string following each prefix is ​​an element in the projected dataset corresponding to that prefix. For example, for the attribute sequence aBbCad, which includes two identical attribute identifiers "a," when obtaining the projected dataset corresponding to each prefix, the attribute identifier strings following the prefix in the projected dataset include "BbCad" and "d."

[0059] S122: Looping through a prefix in the currently acquired projection data set, combining the traversed prefix with a designated prefix corresponding to the currently acquired projection data set to obtain a new designated prefix, and obtaining a projection data set corresponding to the new designated prefix, until the projection data set corresponding to the new designated prefix is ​​an empty set, wherein the projection data set corresponding to the new designated prefix includes an attribute identification string in the currently acquired projection data set that follows the traversed prefix.

[0060] In this embodiment, a prefix obtained by traversal is combined with a designated prefix corresponding to the currently obtained projection dataset, that is, a prefix is ​​sequentially arranged after the designated prefix corresponding to the currently obtained projection dataset, thereby obtaining a new designated prefix.

[0061] S123: Using the specified prefix corresponding to the empty set as the attribute sequence pattern feature associated with the attribute of the target object.

[0062] See Figure 4 , Figure 4 yes Figure 3 The flowchart of step S121 in the embodiment shown is an exemplary embodiment. In step S121, traversing a prefix in all attribute sequences may include the following steps:

[0063] Step S1211: Acquire attribute identifiers that appear in all attribute sequences.

[0064] Step S1212: Determine the number of attribute sequences containing corresponding attribute identifiers.

[0065] Step S1213: Determine the attribute identifiers whose support is greater than a support threshold as a prefix, where the support threshold is proportional to the number of attribute sequences.

[0066] In this embodiment, the support threshold = the number of attribute sequences * the minimum support rate, where the minimum support rate is a number greater than or equal to zero and less than or equal to 1, and the minimum support rate is, for example, 0.4.

[0067] The support of an attribute identifier is the number of attribute sequences containing the corresponding attribute identifier. An attribute identifier whose support is greater than the support threshold is an attribute identifier whose number of attribute sequences containing the corresponding attribute identifier is greater than the support threshold.

[0068] Through the above method, all attribute identifiers that meet the support threshold can be screened out, and the corresponding attribute identifiers are determined as a prefix.

[0069] See Figure 5 , Figure 5 yes Figure 3 In the illustrated embodiment, step S122 is a flow chart of an exemplary embodiment. In step S122, traversing a prefix in the currently acquired projection data set may include the following steps:

[0070] Step S1221: Acquire the attribute identifiers that appear in the currently acquired projection dataset.

[0071] Step S1222: Determine the number of attribute identification strings containing corresponding attribute identifications in the currently acquired projection data set.

[0072] Step S1223: Determine the attribute identifiers whose support is greater than a support threshold as a prefix, where the support threshold is proportional to the number of attribute identifier strings.

[0073] In this step, the support threshold is equal to the number of attribute identifier strings in the projected dataset * the minimum support rate. The minimum support rate is a pre-set value greater than or equal to zero and less than or equal to 1.

[0074] The support of an attribute identifier is the number of attribute identifier strings containing the corresponding attribute identifier. An attribute identifier with support greater than a support threshold is an attribute identifier with the number of attribute identifier strings containing the corresponding attribute identifier greater than the support threshold.

[0075] In this embodiment, the currently acquired projection data set is taken as the processing object, and a prefix traversal is performed on the projection data set. If there are multiple currently acquired projection data sets, a prefix traversal is performed on each projection data set respectively.

[0076] In a specific embodiment, the process of performing sequence pattern mining on multiple attribute sequences of a virtual object of target object A to obtain attribute sequence pattern features associated with the attributes of target object A is as follows:

[0077] Let's assume that the multiple attribute sequences of target object A for the virtual object are as shown in Table 1:

[0078] Attribute Sequence AaBbC AaBcA AcBbd

[0079] Table 1

[0080] First, perform a prefix traversal on multiple attribute sequences.

[0081] Specifically, the attribute identifiers "A", "B", "C", "a", "b", "c", and "d" appearing in multiple attribute sequences are obtained.

[0082] Determine the number of attribute sequences containing the corresponding attribute identifier. The number of attribute sequences containing the attribute identifier "A" is 3, the number of attribute sequences containing the attribute identifier "B" is 3, the number of attribute sequences containing the attribute identifier "C" is 1, the number of attribute sequences containing the attribute identifier "a" is 2, the number of attribute sequences containing the attribute identifier "b" is 2, the number of attribute sequences containing the attribute identifier "c" is 2, and the number of attribute sequences containing the attribute identifier "d" is 1.

[0083] The attribute identifiers whose support is greater than the support threshold are determined as a prefix.

[0084] In this step, let's assume the minimum support rate is 0.4, so the support threshold is 1.2, which means that the number of attribute sequences containing the corresponding attribute identifier must be at least 2 to be a prefix. Therefore, a prefix specifies the prefix including "A", "B", "a", "b", and "c".

[0085] Get the projection dataset corresponding to the specified prefix, as shown in Table 2:

[0086]

[0087]

[0088] Table 2

[0089] Taking the specified prefix "A" as an example, the step of traversing a prefix in the currently acquired projection dataset is described in detail, that is, traversing a prefix in the projection dataset corresponding to the specified prefix "A".

[0090] Get the attribute identifiers that appear in the currently acquired projection dataset, "a", "B", "b", "C", "c", "d", "A".

[0091] Determine the number of attribute identification strings containing corresponding attribute identifications in the currently acquired projection data set, the number of attribute identification strings containing attribute identification "a" is 2, the number of attribute identification strings containing attribute identification "B" is 3, the number of attribute identification strings containing attribute identification "b" is 2, the number of attribute identification strings containing attribute identification "C" is 1, the number of attribute identification strings containing attribute identification "c" is 2, the number of attribute identification strings containing attribute identification "d" is 1, and the number of attribute identification strings containing attribute identification "A" is 1.

[0092] The attribute identifiers whose support is greater than the support threshold are determined as a prefix, and the support threshold is proportional to the number of attribute identifier strings.

[0093] The minimum support rate is still set to 0.4, and the support threshold is still 1.2, that is, the number of attribute identifier strings containing the corresponding attribute identifier must be at least 2 to serve as a prefix. Therefore, a prefix includes "B", "a", "b", and "c".

[0094] Since the designated prefix corresponding to the currently acquired projection dataset is "A", a prefix obtained through traversal is combined with the designated prefix corresponding to the currently acquired projection dataset to obtain new designated prefixes of "AB", "Aa", "Ab", and "Ac".

[0095] Get the projection dataset corresponding to the new specified prefix, as shown in Table 3:

[0096]

[0097]

[0098] Table 3

[0099] The projection data set corresponding to the newly specified prefix "AB" is used as the currently acquired projection data set to execute traversing a prefix in the currently acquired projection data set.

[0100] Get the attribute identifiers that appear in the currently acquired projection dataset, "b", "C", "c", "d", and "A".

[0101] Attribute identifiers with support greater than the support threshold are considered a prefix. The support threshold is proportional to the number of attribute identifier strings. If the minimum support is still set to 0.4, the support threshold is still 1.2. To qualify as a prefix, the number of attribute identifier strings containing the corresponding attribute identifier must be at least 2. Therefore, a prefix includes "b."

[0102] Since the designated prefix corresponding to the currently acquired projection dataset is "AB", a prefix obtained by traversal is combined with the designated prefix corresponding to the currently acquired projection dataset to obtain a new designated prefix "ABb".

[0103] Get the projection dataset corresponding to the new specified prefix, as shown in Table 4:

[0104] Specify a prefix Corresponding suffix ABb C

[0105] Table 4

[0106] Traversing a prefix in the currently acquired projection data set, let's assume that the minimum support rate is 0.4. Since the attribute identifier string containing the attribute identifier C is 1, the attribute identifier C is a prefix.

[0107] A prefix is ​​combined with the specified prefix to obtain a new specified prefix ABbC. Since the projection data set corresponding to the new specified prefix is ​​an empty set, the loop end condition is met. Therefore, the sequence ABbC is an attribute sequence pattern feature of the target object A.

[0108] Through the above method, multiple attribute sequence pattern features that are deeply associated with the attributes of the target object can be mined.

[0109] See Figure 6 , Figure 6 yes Figure 1 The flowchart of step S200 in the embodiment shown is an exemplary embodiment. Step S200 may include the following steps:

[0110] Step S210: Determine the probability of the target object preferring the virtual object, and the mutual information between event A and event Q. In this embodiment, let event A be the target object's preference for the virtual object, and event Q be the occurrence of the fused feature. The fused feature includes mutually independent target object features, virtual object features, and attribute sequence pattern features associated with the target object's attribute sequence for the corresponding virtual object.

[0111] A sample is obtained, where the sample includes the target object's preference for all virtual objects. The preference indicates whether the target object has a preference for the virtual object, and the preference can be a preference or a non-preference.

[0112] The probability that the target object prefers the virtual object is equal to the total number of samples of the virtual object preferred by the target object / the total number of virtual objects.

[0113] Step S220: Calculate the corresponding exponential function value with 10 as the base and the mutual information between event A and event Q as the exponent.

[0114] That is, the exponential function value = 10 I(Q，A) .

[0115] Where I(Q, A) represents the mutual information between event A and event Q.

[0116] Let's assume that the occurrence of the target object feature is event q1, the occurrence of the corresponding virtual object feature is event q2, and the occurrence of the attribute sequence pattern feature associated with the target object's attribute sequence for the corresponding virtual object is event q3. Since q1, q2, and q3 are independent of each other, we have:

[0117] I(Q,A)=I(q1,A)+I(q2,A)+I(q3,A).

[0118] Among them, I(q1, A) represents the mutual information between event q1 and event Q, I(q2, A) represents the mutual information between event q2 and event Q, and I(q3, A) represents the mutual information between event q3 and event Q.

[0119] Step S230: Calculate the product of the exponential function value and the preference probability to obtain the recommendation probability.

[0120] In summary, in the technical solutions provided in the embodiments of this application,

[0121] After obtaining the attribute sequences of the same target object for different virtual objects, this embodiment performs sequence mining on all attribute sequences, mining attribute subsequences with an occurrence frequency greater than a certain threshold. These subsequences are used as attribute sequence pattern features that can reflect the target object's preferences for different virtual objects. Furthermore, using the fusion features obtained by fusing the target object features, the virtual object features corresponding to multiple virtual objects to be recommended, and the attribute sequence pattern features related to the target object, the probability of recommending the virtual object to the target object is calculated, and virtual objects with a recommendation probability greater than a preset threshold are recommended to the target object. The information recommendation method provided by this embodiment can mine effective information that can reflect the target object's attribute preferences from the target object's multiple attribute sequence data for multiple virtual objects, thereby improving the accuracy of information recommendation.

[0122] Figure 7 FIG. 1 is a block diagram of an information recommendation device according to an exemplary embodiment of the present application. Figure 7 As shown, the device includes:

[0123] The feature acquisition module 510 is used for multiple recommendation features, which include target object features, virtual object features corresponding to multiple virtual objects to be recommended, and attribute sequence pattern features associated with the attribute sequence of the target object, wherein the attribute sequence of the target object includes the attribute sequences of the target object for different virtual objects respectively, and the attribute sequence includes multiple attribute identifiers arranged in chronological order, and the attribute identifiers are used to represent the historical operation attribute types of the target object for the corresponding virtual objects.

[0124] The probability acquisition module 520 is connected to the feature acquisition module and is configured to fuse a plurality of recommendation features to obtain a fused feature, and determine the recommendation probability of the virtual object recommended to the target object based on the fused feature.

[0125] The recommendation module 530 is connected to the probability acquisition module and is used to recommend virtual objects with a recommendation probability greater than a preset threshold to the target object.

[0126] In another exemplary embodiment, the feature acquisition module 510 includes:

[0127] The attribute sequence acquisition unit is used to acquire the attribute sequences of the target object for different virtual objects.

[0128] The mining unit is connected to the attribute sequence acquisition unit and is used to perform sequence pattern mining on all acquired attribute sequences to obtain attribute sequence pattern features associated with the attributes of the target object.

[0129] In another exemplary embodiment, the excavation unit includes:

[0130] The attribute sequence traversal subunit is used to traverse a prefix in all attribute sequences, take a prefix as a specified prefix, and obtain the projection data set corresponding to the specified prefix. The projection data set corresponding to the specified prefix includes the attribute identification string located after the corresponding prefix in all attribute sequences.

[0131] The attribute identification string traversal subunit is connected to the attribute sequence traversal subunit and is used to cyclically execute the steps of traversing a prefix in the currently acquired projection data set, combining the traversed prefix with the specified prefix corresponding to the currently acquired projection data set to obtain a new specified prefix, and obtaining the projection data set corresponding to the new specified prefix, until the result data set corresponding to the new specified prefix is ​​an empty set, wherein the projection data set corresponding to the new specified prefix includes the attribute identification string in the currently acquired projection data set that is located after the traversed prefix.

[0132] The result acquisition subunit is connected to the attribute identification string traversal subunit, and is used to use the specified prefix corresponding to the empty set as the attribute sequence pattern feature associated with the attribute of the target object.

[0133] In another exemplary embodiment, the attribute sequence traversal subunit includes:

[0134] The first attribute identifier acquisition subunit is used to acquire attribute identifiers that appear in all attribute sequences.

[0135] The first determining subunit is connected to the first attribute identifier obtaining subunit and is used to determine the number of attribute sequences containing corresponding attribute identifiers.

[0136] The second determining subunit is connected to the first determining subunit and is configured to determine an attribute identifier whose support is greater than a support threshold as a prefix, where the support threshold is proportional to the number of attribute sequences.

[0137] In another exemplary embodiment, the attribute identification string traversal subunit includes:

[0138] The second attribute identifier acquisition subunit is used to acquire the attribute identifier appearing in the currently acquired projection data set.

[0139] The third determining subunit is connected to the second attribute identifier obtaining subunit and is used to determine the number of attribute identifier strings containing the corresponding attribute identifier in the currently obtained projection data set.

[0140] The fourth determining subunit is connected to the third determining subunit and is used to determine the attribute identifier whose support is greater than a support threshold as a prefix, and the support threshold is proportional to the number of attribute identifier strings.

[0141] In another exemplary embodiment, the probability acquisition module 520 includes:

[0142] The determination unit is used to determine the target object's preference probability for the virtual object and the mutual information between event A and event Q. The target object's preference for the virtual object is event A, and the occurrence of the fusion feature is event Q.

[0143] The first calculation unit is connected to the determination unit and is used to calculate a corresponding exponential function value with 10 as the base and the mutual information as the exponent.

[0144] The second calculation unit is connected to the first calculation unit and is used to calculate the product of the exponential function value and the preference probability to obtain the recommendation probability.

[0145] It should be noted that the apparatus provided in the above embodiment and the method provided in the above embodiment belong to the same concept, wherein the specific manner in which each module and unit performs operations has been described in detail in the method embodiment and will not be repeated here.

[0146] An embodiment of the present application further provides an electronic device, including a processor and a memory, wherein the memory stores computer-readable instructions, which implement the above information recommendation method when executed by the processor.

[0147] Figure 8 A schematic diagram of the structure of a computer system suitable for implementing an electronic device according to an embodiment of the present application is shown.

[0148] It should be noted that Figure 8 The computer system 1000 of the electronic device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0149] like Figure 8 As shown, the computer system 1000 includes a central processing unit (CPU) 1001, which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 1002 or the program loaded from the storage part 1008 to the random access memory (RAM) 1003, such as executing the information recommendation method described in the above embodiment. Various programs and data required for system operation are also stored in the RAM 1003. The CPU 1001, ROM 1002, and RAM 1003 are connected to each other via a bus 1004. An input / output (I / O) interface 1005 is also connected to the bus 1004.

[0150] The following components are connected to the I / O interface 1005: an input section 1006 including a keyboard, a mouse, and the like; an output section 1007 including devices such as a cathode ray tube (CRT), a liquid crystal display (LCD), and a speaker; a storage section 1008 including a hard disk; and a communication section 1009 including a network interface card such as a LAN (Local Area Network) card or a modem. The communication section 1009 performs communication processing via a network such as the Internet. A drive 1010 is also connected to the I / O interface 1005 as needed. Removable media 1011, such as a magnetic disk, an optical disk, a magneto-optical disk, or a semiconductor memory, is installed in the drive 1010 as needed, so that computer programs read therefrom can be installed into the storage section 1008 as needed.

[0151] In particular, according to an embodiment of the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present application includes a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a computer program for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network via the communication section 1009, and / or installed from a removable medium 1011. When the computer program is executed by the central processing unit (CPU) 1001, the various functions defined in the system of the present application are executed.

[0152] It should be noted that the computer-readable medium shown in the embodiments of the present application may be a computer-readable signal medium or a computer-readable storage medium or any combination of the two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM), a flash memory, an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, device or device. In the present application, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, which carries a computer-readable computer program. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium that can transmit, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. A computer program embodied on a computer-readable medium may be transmitted using any suitable medium, including but not limited to wireless, wired, or any suitable combination thereof.

[0153] The flowcharts and block diagrams in the accompanying drawings illustrate the possible implementation architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present application. Among them, each box in the flowchart or block diagram can represent a module, program segment, or part of the code, and the above-mentioned module, program segment, or part of the code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in an order different from that marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flowchart, and the combination of boxes in the block diagram or flowchart, can be implemented with a dedicated hardware-based system that performs the specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0154] The units involved in the embodiments described in this application may be implemented by software or hardware, and the units described may also be set in a processor. In some cases, the names of these units do not constitute limitations on the units themselves.

[0155] Another aspect of the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the aforementioned information recommendation method. The computer-readable storage medium may be included in the electronic device described in the above embodiments, or may exist independently and not be incorporated into the electronic device.

[0156] Another aspect of the present application provides a computer program product or computer program, which includes computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the information recommendation method provided in each of the above embodiments.

[0157] The above content is only a preferred exemplary embodiment of the present application and is not intended to limit the implementation scheme of the present application. Ordinary technicians in this field can easily make corresponding changes or modifications based on the main ideas and spirit of the present application. Therefore, the scope of protection of the present application shall be based on the scope of protection required by the claims.

Claims

1. An information recommendation method, characterized in that: include: Acquiring a plurality of recommendation features, the plurality of recommendation features comprising a target object feature, virtual object features corresponding to a plurality of virtual objects to be recommended, and an attribute sequence pattern feature associated with an attribute sequence of the target object, wherein the attribute sequence of the target object comprises attribute sequences of the target object for different virtual objects, the attribute sequence comprising a plurality of attribute identifiers arranged in chronological order, the attribute identifiers being used to represent attribute types of historical operations of the target object with respect to the corresponding virtual objects; fusing the plurality of recommendation features to obtain a fused feature, and determining a recommendation probability of recommending the virtual object to the target object based on the fused feature; Recommending a virtual object whose recommendation probability is greater than a preset threshold to the target object; The determining, based on the fusion feature, of the recommendation probability of recommending the virtual object to the target object includes: Determining a probability of the target object preferring the virtual object, and mutual information between a first event and a second event, wherein the first event is used to indicate that the target object prefers the virtual object, and the second event is used to indicate that the fusion feature has occurred; Calculating an exponential function value corresponding to the mutual information; The product of the exponential function value and the preference probability is calculated to obtain the recommendation probability.

2. The method according to claim 1, characterized in that The obtaining of multiple recommendation features includes: Obtaining attribute sequences of the target object for different virtual objects respectively; Sequence pattern mining is performed on all acquired attribute sequences to obtain attribute sequence pattern features associated with the attributes of the target object.

3. The method according to claim 2, characterized in that The performing sequence pattern mining on the attribute sequence to obtain attribute sequence pattern features associated with the attributes of the target object includes: Traversing a prefix in all attribute sequences, taking the prefix as a designated prefix, and obtaining a projection data set corresponding to the designated prefix, where the projection data set corresponding to the designated prefix includes an attribute identification string located after the corresponding prefix in all attribute sequences; Looping through a prefix in the currently acquired projected data set, combining the traversed prefix with a specified prefix corresponding to the currently acquired projected data set to obtain a new specified prefix, and obtaining a projected data set corresponding to the new specified prefix, until the result data set corresponding to the new specified prefix is ​​an empty set, wherein the projected data set corresponding to the new specified prefix includes the attribute identification string in the currently acquired projected data set that follows the traversed prefix; The specified prefix corresponding to the empty set is used as an attribute sequence pattern feature associated with the attribute of the target object.

4. The method according to claim 3, characterized in that The method traverses a prefix in all attribute sequences, including: Get the attribute identifiers that appear in all the attribute sequences; Determine the number of attribute sequences containing the corresponding attribute identifier; An attribute identifier whose support is greater than a support threshold is determined as the prefix, and the support threshold is proportional to the number of the attribute sequences.

5. The method according to claim 3, characterized in that The step of traversing a prefix in the currently acquired projection dataset includes: Obtaining attribute identifiers that appear in the currently acquired projection dataset; Determining the number of attribute identification strings containing corresponding attribute identifications in the currently acquired projection data set; An attribute identifier with a support greater than a support threshold is determined as the prefix, where the support threshold is proportional to the number of the attribute identifier strings.

6. The method according to claim 1, wherein The obtaining of multiple recommendation features includes: The target object features and the virtual object features are preprocessed, where the preprocessing includes at least one of missing value processing and outlier processing.

7. An information recommendation device, characterized in that: include: a feature acquisition module, configured to be used for multiple recommendation features, the multiple recommendation features including target object features, virtual object features corresponding to multiple virtual objects to be recommended, and attribute sequence pattern features associated with an attribute sequence of the target object, wherein the attribute sequence of the target object includes attribute sequences of the target object for different virtual objects, the attribute sequence including multiple attribute identifiers arranged in chronological order, the attribute identifiers being used to represent attribute types of historical operations of the target object with respect to the corresponding virtual objects; a probability acquisition module, connected to the feature acquisition module, configured to fuse the multiple recommendation features to obtain a fusion feature, and determine a recommendation probability of the virtual object recommended to the target object based on the fusion feature; a recommendation module, connected to the probability acquisition module, configured to recommend to the target object a virtual object whose recommendation probability is greater than a preset threshold; The probability acquisition module is further configured to perform the following steps: Determining a probability of the target object preferring the virtual object, and mutual information between a first event and a second event, wherein the first event is used to indicate that the target object prefers the virtual object, and the second event is used to indicate that the fusion feature has occurred; Calculating an exponential function value corresponding to the mutual information; The product of the exponential function value and the preference probability is calculated to obtain the recommendation probability.

8. An electronic device, characterized in that: include: a memory storing computer-readable instructions; A processor reads computer-readable instructions stored in the memory to execute the method according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that Computer-readable instructions are stored thereon, and when the computer-readable instructions are executed by a processor of a computer, the computer is caused to execute the method according to any one of claims 1 to 6.

10. A computer program product, characterized in that comprising computer instructions stored in a computer-readable storage medium; A processor of a computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device executes the method according to any one of claims 1 to 6.

Images