Virtual community user behavior prediction method and device based on case reasoning

Abstract

The invention relates to a case reasoning-based virtual community user behavior prediction method and device. The method comprises the following steps: acquiring a user behavior case in a virtual community according to a preset virtual community; clustering the user behavior cases based on information entropy increase, and constructing a source case library according to a clustering result of the user behavior cases; obtaining a to-be-predicted target case solution, and retrieving the source case library according to the target case, the source case library and a preset case recommendation threshold; when the retrieval case set is not empty, cases in the retrieval case set are reused, when the retrieval case set is empty, the source case is revised, and whether the revised source case is reserved or not is determined based on the similarity between the revised case and the most similar case in the source case library; and predicting the user behavior based on the reused source case. By adopting the method, accurate prediction of user behaviors can be realized.

Description

Technical Field

[0001] The present application relates to the technical field of virtual community user behavior prediction, and in particular to a method and device for virtual community user behavior prediction based on case-based reasoning. Background Art

[0002] Virtual communities are online social communities composed of people with shared interests and needs, including knowledge question-and-answer communities, online education communities, and e-commerce platforms. These widespread online communities provide platforms for participating members to share information and solve problems. Community incentive mechanisms can guide and regulate the behavior of community users and their interactions. Therefore, incentive mechanisms play a vital role in maintaining the stable development and prosperity of a community. Predicting community user behavior after adjustments to incentive mechanisms can help community management verify whether these incentive mechanisms are conducive to community stability and prosperity, and aid in making sound decisions. However, predicting community user behavior is a challenging task, involving multiple disciplines, including psychology, behavioral science, sociology, and computer science.

[0003] Some studies have explored various motivations and factors that influence community user behavior, including entertainment, ideology, prestige, reciprocity, efficacy, attachment, and community resources. Other studies have explored the potential impact of these factors on community user behavior. However, existing research remains uncertain when it comes to why some community users engage in behaviors that others do not. In fact, the only reliable analysis seems to rely on historical information about community users' behavior. Summary of the Invention

[0004] Based on this, it is necessary to provide a method and device for predicting user behavior in a virtual community based on case-based reasoning to address the above technical issues.

[0005] A method for predicting user behavior in a virtual community based on case-based reasoning, the method comprising:

[0006] According to a pre-set virtual community, obtaining user behavior cases in the virtual community; the user behavior cases are represented by user behavior attributes and user behavior types;

[0007] Clustering the user behavior cases based on information entropy increase, and building a source case library according to the clustering results of the user behavior cases;

[0008] Obtain a solution to a target case to be predicted, and search a source case library based on the target case, the case library, and a preset case recommendation threshold;

[0009] When the retrieval case set is not empty, the cases in the retrieval case set are reused; when the retrieval case set is empty, the source case is revised, and based on the similarity between the revised case and the most similar case in the case library, it is determined whether to retain the revised source case;

[0010] Predict user behavior based on reused source cases.

[0011] In one embodiment, the method further includes: obtaining user behavior cases cb= in the virtual community according to a preset virtual community.<u,b> ;

[0012] Where u=<rew,pr,pc,op> , rew represents the interval number of the community user incentive value, pr represents the preference level of the community user behavior, pc represents the preference type of the community user behavior, op represents the relational operator, and b represents the community user behavior type; the community user behavior types include: proposing task behavior, solution behavior, like behavior, and dislike behavior.

[0013] In one embodiment, the method further includes: obtaining the information entropy of the j-th behavior case of the i-th community user as follows:

[0014]

[0015] Among them, case CB i,j Probability CB i，j represents the number of occurrences of the jth type of behavior of the i-th community user, i∈{1..n}, n is the number of community users in the case library, j∈{0,1,2,3}, representing task behavior, solution behavior, like behavior, and dislike behavior respectively;

[0016] Calculate the information entropy of the j-th behavior case based on the information entropy of the j-th behavior case of the i-th community user for:

[0017]

[0018] According to the information entropy of the j-th behavior case The information entropy of the j-th behavior case of the i-th community user is calculated to obtain the information entropy of the k-th case cluster of the j-th behavior for:

[0019]

[0020] According to the information entropy of the k-th case cluster of the j-th behavior Calculate the average entropy of the k-th case cluster of the j-th behavior for:

[0021]

[0022] According to the information entropy of the j-th behavior case Calculate the entropy increase of the k-th case cluster of the j-th behavior as:

[0023]

[0024] When clustering user behavior cases, obtain the community user behavior data SD and the behavior category set B, calculate the behavior case probability of each community user, and select the community user maxP with the largest probability in behavior b as the cluster center SD maxP Put it into the cluster, select the community user case that is most similar to the community user maxP behavior case in the community user behavior data SD based on the principle of information entropy increase, and add it to the cluster until the average entropy of the cluster no longer increases, and move the cluster center SD maxP Add to the case base CB and delete the cluster cases that have been added to the cluster set from the community user behavior data SD.

[0025] In one embodiment, the method further includes: obtaining the target case, the case library, and a preset case recommendation threshold;

[0026] The similarity function is constructed as:

[0027]

[0028] Among them, u and v represent cases, u i and v i represents the i-th element of the two cases, λ i Represents the similarity calculation weight of the i-th element;

[0029] According to a similarity function, the similarity between the target case and the retrieval case in the source case library is compared, and when the similarity is greater than a case recommendation threshold, the retrieval case is reused.

[0030] In one embodiment, the further step includes: constructing a revision function as follows:

[0031]

[0032] Where t is the target case, t rew represents the incentive value of the target case t, s represents the revised case after revision, u and v represent the adjacent cases respectively;

[0033] The source case is revised according to the revision function.

[0034] In one embodiment, the method further includes: when the similarity between the revised case and the most similar case in the source case library is less than or equal to a threshold, the revised case is retained; when the similarity between the revised case and the most similar case in the source case library is greater than the threshold, the revised case is not retained.

[0035] In one embodiment, the method further includes evaluating the predicted user behavior.

[0036] A virtual community user behavior prediction device based on case-based reasoning, the device comprising:

[0037] A user behavior case acquisition module is used to acquire user behavior cases in a pre-set virtual community according to the virtual community; the user behavior cases are represented by user behavior attributes and user behavior types;

[0038] A case library construction module is used to cluster the user behavior cases based on information entropy increase, and to construct a source case library according to the clustering results of the user behavior cases;

[0039] A retrieval module is used to obtain a solution to a target case to be predicted, and retrieve the source case library according to the target case, the source case library, and a preset case recommendation threshold;

[0040] a case processing module, configured to reuse a case in the retrieved case set when the retrieved case set is not empty, and revise a source case when the retrieved case set is empty, and determine whether to retain the revised case based on a similarity between the revised case and the most similar case in the source case library;

[0041] The behavior prediction module is used to predict user behavior based on reused source cases.

[0042] A computer device includes a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0043] According to a pre-set virtual community, obtaining user behavior cases in the virtual community; the user behavior cases are represented by user behavior attributes and user behavior types;

[0044] Clustering the user behavior cases based on information entropy increase, and building a source case library according to the clustering results of the user behavior cases;

[0045] Obtain a solution to a target case to be predicted, and search the source case library according to the target case, the source case library, and a preset case recommendation threshold;

[0046] When the retrieval case set is not empty, the cases in the retrieval case set are reused; when the retrieval case set is empty, the source case is revised, and based on the similarity between the revised case and the most similar case in the case library, it is determined whether to retain the revised source case;

[0047] Predict user behavior based on reused source cases.

[0048] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the following steps:

[0049] According to a pre-set virtual community, obtaining user behavior cases in the virtual community; the user behavior cases are represented by user behavior attributes and user behavior types;

[0050] Clustering the user behavior cases based on information entropy increase, and building a source case library according to the clustering results of the user behavior cases;

[0051] Obtain a solution to a target case to be predicted, and search the source case library according to the target case, the source case library, and a preset case recommendation threshold;

[0052] When the retrieval case set is not empty, the cases in the retrieval case set are reused; when the retrieval case set is empty, the source case is revised, and based on the similarity between the revised case and the most similar case in the case library, it is determined whether to retain the revised source case;

[0053] Predict user behavior based on reused source cases.

[0054] The above-mentioned virtual community user behavior prediction method and device based on case reasoning describes the diverse behavior patterns of users by constructing a case library, and designs a user behavior case reasoning algorithm to realize dynamic adjustment of the case library, so that it can change according to changes in user behavior scenarios during the prediction process, and use the revised user behavior cases to predict user behaviors in different behavior scenarios from the past, thereby improving the accuracy of user behavior prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 1 is a flow chart of a method for predicting user behavior in a virtual community based on case-based reasoning in one embodiment;

[0056] Figure 2 Schematic diagram of pseudo code of a community user behavior case clustering algorithm based on entropy increase in one embodiment;

[0057] Figure 3 Schematic diagram of the steps of community user behavior inference in one embodiment;

[0058] Figure 4 Schematic diagram of a community user behavior inference algorithm in one embodiment;

[0059] Figure 5 is a schematic diagram of a retrieval algorithm in one embodiment;

[0060] Figure 6 A schematic diagram of a modification algorithm in one embodiment;

[0061] Figure 7 A schematic diagram of a retention algorithm in one embodiment;

[0062] Figure 8 1 is a structural block diagram of a virtual community user behavior prediction device based on case-based reasoning in one embodiment;

[0063] Figure 9 FIG. 1 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0064] In order to make the purpose, technical solutions and advantages of this application more clear, the following further describes this application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0065] In one embodiment, Figure 1 As shown, a virtual community user behavior prediction method based on case-based reasoning is provided, which includes the following steps:

[0066] Step 102: Obtain user behavior cases in the virtual community according to the preset virtual community.

[0067] User behavior cases are represented by user behavior attributes and user behavior types.

[0068] Step 104 : clustering the user behavior cases based on information entropy increase, and building a source case library according to the clustering results of the user behavior cases.

[0069] Step 106 , obtaining a solution to the target case to be predicted, and searching the source case library according to the target case, the source case library, and a preset case recommendation threshold.

[0070] Step 108: When the retrieval case set is not empty, the cases in the retrieval case set are reused; when the retrieval case set is empty, the source case is revised, and based on the similarity between the revised case and the most similar case in the case library, it is determined whether to retain the revised source case.

[0071] Step 110 : predicting user behavior based on the reused source cases.

[0072] In the above-mentioned virtual community user behavior prediction method based on case reasoning, a case library is constructed to describe the diverse behavior patterns of users, and a user behavior case reasoning algorithm is designed to realize the dynamic adjustment of the case library, so that it can change according to the changes in user behavior scenarios during the prediction process. The revised user behavior cases are used to predict user behaviors in different behavior scenarios from the past, thereby improving the accuracy of user behavior prediction.

[0073] In one embodiment, based on a pre-set virtual community, a user behavior case cb=<u,b> ; where u=<rew,pr,pc,op> , rew represents the interval number of community user incentive value, pr represents the preference level of community user behavior, pc represents the preference type of community user behavior, op represents the relational operator, and b represents the community user behavior type; community user behavior types include: proposing task behavior, solution behavior, like behavior, and dislike behavior.

[0074] Specifically, describing community user behavior cases is a crucial step in building a case library. Selecting too many parameters to describe community user behavior scenarios will lead to complex calculations and collinearity issues; conversely, selecting too few parameters will not guarantee the accuracy of behavioral reasoning. According to empirical analysis, community users' behavioral decisions are related to their inherent attributes, task-related information (such as the number of solutions to a task), and incentive mechanisms. However, within a specific time period, community users can always find tasks that are suitable for them to solve, and the incentive mechanism remains unchanged during that time period. Therefore, community user behavior decision scenarios are described only through community user attribute characteristics and behavior types, as shown in Table 1.

[0075] Table 1 Element description of community user behavior cases

[0076]

[0077] In addition, in order to obtain the decision-making behavior patterns of community users in different scenarios, it is necessary to cluster the behavioral cases of community users and select representative behavioral cases to build a source case library.

[0078] Existing research differs from existing clustering methods, such as K-means, kin propagation (AP), and DBSCAN. These algorithms share a common characteristic: they require a predefined number of clusters or specific parameters to complete the classification. For example, the number of clusters (K) and the starting center point in K-means, and the preference and damping factor in kin propagation. These parameter values ​​are determined based on expert experience or prior information, resulting in subjectivity and poor adaptability. Given the large number of community users, the number of behavioral decision patterns, and parameters such as cluster centers, it is difficult to determine these parameters empirically. To address this issue, this paper proposes a method based on information entropy increase that selects the case with the highest information entropy as the initial cluster center. By combining entropy increase with case similarity, the number of clusters is automatically determined.

[0079] In one embodiment, the information entropy of the j-th behavior case of the i-th community user is obtained as:

[0080]

[0081] Among them, case CB i,j Probability CB i，j represents the number of occurrences of the jth type of behavior of the i-th community user, i∈{1..n}, n is the number of community users in the case library, j∈{0, 1, 2, 3}, representing task behavior, solution behavior, like behavior, and dislike behavior respectively;

[0082] Calculate the information entropy of the j-th behavior case based on the information entropy of the j-th behavior case of the i-th community user for:

[0083]

[0084] According to the information entropy of the j-th behavior case The information entropy of the j-th behavior case of the i-th community user is calculated to obtain the information entropy of the k-th case cluster of the j-th behavior for:

[0085]

[0086] According to the information entropy of the k-th case cluster of the j-th behavior Calculate the average entropy of the k-th case cluster of the j-th behavior for:

[0087]

[0088] According to the information entropy of the j-th behavior case Calculate the entropy increase of the k-th case cluster of the j-th behavior as:

[0089]

[0090] When clustering user behavior cases, obtain the community user behavior data SD and the behavior category set B, calculate the behavior case probability of each community user, and select the community user maxP with the largest probability in behavior b as the cluster center SD maxP Put it into the cluster, select the community user case that is most similar to the community user maxP behavior case in the community user behavior data SD based on the principle of information entropy increase, and add it to the cluster until the average entropy of the cluster no longer increases, and move the cluster center SD maxP Add to the case base CB and delete the cluster cases that have been added to the cluster set from the community user behavior data SD.

[0091] Specifically, the pseudo code of the community user behavior case clustering algorithm based on entropy increase is as follows: Figure 2 As shown, in Figure 2 The following steps are included.

[0092] 1) Calculate the behavior case probability. The first line calculates the behavior case probability of each community user based on the behavior category.

[0093] 2) Initialize relevant clustering variables. Lines 4-7 initialize relevant clustering variables. Line 4 selects the community user maxP with the highest probability in behavior b, and line 7 places it as the cluster center SDmaxP in the cluster.

[0094] 3) Clustering community user behavior cases. Lines 8-23 select the community user cases in SD that are most similar to the maxP behavior cases based on the principle of entropy increase and add them to the cluster until the average entropy of the cluster no longer increases. Lines 24-25 add the cluster center SDmaxP to the case base CB. Cluster cases that have already been added to the cluster set are deleted from SD.

[0095] 4) Return the result. Line 28 returns the case library CB.

[0096] In one embodiment, the community user behavior inference step is as follows: Figure 3 As shown in Figure 1, the community user behavior reasoning algorithm dynamically adjusts the community user behavior case library, and the changes in community user behavior patterns after the adaptive incentive mechanism is adjusted. The community user behavior reasoning process includes four steps: case retrieval, case reuse, case revision, and case retention. The community user behavior reasoning algorithm BR (Behavior Reasoning) is as follows: Figure 4 shown.

[0097] exist Figure 4In [1], the algorithm inputs are: target case tcase that needs a solution, case base CB, case recommendation threshold δ, and case retention threshold ζ. The algorithm output is: recommended source case solvedCase.

[0098] The algorithm includes the following steps.

[0099] 1) Case retrieval. Line 1 retrieves the appropriate source case scase based on the target case tcase, the case base CB, and the case recommendation threshold.

[0100] 2) Reuse case. Lines 2-3 indicate that if scase is not empty, it will be reused as the recommended source case solvedCase to predict community user behavior.

[0101] 3) Revise and retain the case. Lines 4-7 indicate that scase is empty, meaning that previous behavior scenarios cannot be used to infer community user behavior. At this point, adaptive adjustments are required to revise the community user behavior case. A threshold ζ is used to determine whether to retain the revised case solvedCase. If it meets the requirements, it is retained as a learned case in the community user behavior case library, enabling dynamic adjustment of the community user behavior case library and supporting current and future predictions of community user behavior in the same scenario.

[0102] 4) Return the result. Line 8 returns the recommended source case solvedCase.

[0103] In one embodiment, a target case, a case library, and a preset case recommendation threshold are obtained; and a similarity function is constructed as follows:

[0104]

[0105] Among them, u and v represent cases, u i and v i represents the i-th element of the two cases, λ i Represents the similarity calculation weight of the i-th element; according to the similarity function, compare the similarity of the target case with the retrieval case in the case library, and when the similarity is greater than the case recommendation threshold, add the retrieval case to the source case set.

[0106] Specifically, the retrieval algorithm is as follows Figure 5 As shown, the algorithm steps are as follows.

[0107] 1) Initialize parameters. Lines 1-2 initialize parameters, including the maximum similarity maxSim and the source case solvedCase recommended as the solution.

[0108] 2) Case retrieval. Lines 3-11 describe the case retrieval process for community users. The function calSimilarity evaluates the similarity between the problem generated by community user behavior (the scenario of the target case) and each source case problem (scenario). The case similarity is calculated using the similarity function. The algorithm compares the source case with the highest similarity. If the similarity sim between this case and the target case tcase is greater than a threshold δ, it is considered a solution and used to infer the contribution behavior of the community user.

[0109] 3) Return the result. Line 12 returns the recommended source case solvedCase.

[0110] Preferably, the case reuse threshold δ is set to 0.5.

[0111] In one embodiment, if a suitable source case is retrieved, it is used as a solution and recommended to an agent representing a community user in a computational model to generate an action.

[0112] In one embodiment, the build revision function is:

[0113]

[0114] Where t is the target case, t rew represents the stimulus value of the target case t, s represents the revised case after revision, u and v represent the adjacent cases respectively; the source case is revised according to the revision function.

[0115] Specifically, when predicting community user behavior, if no suitable case is retrieved, the source case needs to be revised. This situation usually occurs when the external environment such as the adjustment of the incentive mechanism changes and the community user behavior pattern changes. Therefore, no similar community user behavior application scenarios can be retrieved in the community user behavior case library. Therefore, the proposed modification algorithm is as follows Figure 6 The input of the algorithm is the target case tcase and the case base CB. The output is the revised case scase.

[0116] The algorithm includes the following steps.

[0117] 1) Initialize the case. Lines 1-2 initialize the adjacent cases u and v.

[0118] 2) Retrieve adjacent cases. Lines 3-11 search for cases u and v adjacent to the target case t based on the stimulus value.

[0119] 3) Revised case. Lines 12-13 calculate and obtain the revised case s according to the revision function and return it.

[0120] In one embodiment, when the similarity between the revised case and the most similar case of the target case in the case library is less than or equal to a threshold δ, the revised case is retained; when the similarity between the revised case and the most similar case of the target case in the case library is greater than the threshold, the revised case is not retained.

[0121] Specifically, the proposed retention algorithm is as follows Figure 7 As shown in Figure 2, the algorithm inputs are the case base CB, the target case t, the revised case r, and the retention threshold ζ. The output is the revised case base CB.

[0122] The algorithm includes the following steps.

[0123] 1) Initialize parameters. Lines 1-2 initialize parameters, including the maximum similarity maxSim and the source case b.

[0124] 2) Retrieve similar cases. Lines 3-9 describe the retrieval process of the source case b in the case library that is most similar to the target case t.

[0125] 3) Retain Cases. Lines 10-12 describe the case retention process. If the similarity between the most similar source case b and the revised case r is less than the threshold δ, the revised case r is retained in the case base CB.

[0126] 4) Return the result. Line 13 returns the case library that retains the revised case.

[0127] In one embodiment, the predicted user behavior also needs to be evaluated. Specifically, the user behavior prediction results based on case reasoning are compared with the user behavior of the real community to evaluate the effectiveness and universality of the prediction method of this application. For the prediction results of behavior, there are only two values: correct or incorrect. Therefore, this application uses accuracy for evaluation. A higher accuracy rate for community user behavior prediction indicates that the method is effective. The details are as follows:

[0128]

[0129] Among them, TP represents the number of samples predicted correctly, and FP represents the number of samples predicted incorrectly.

[0130] It should be understood that although Figure 1 The steps in the flowchart are shown in sequence as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. In addition, Figure 1At least part of the steps may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least part of the sub-steps or stages of other steps.

[0131] In one embodiment, Figure 8 As shown, a virtual community user behavior prediction device based on case reasoning is provided, including: a user behavior case acquisition module 802, a case library construction module 804, a retrieval module 806, a case processing module 808 and a behavior prediction module 810, wherein:

[0132] A user behavior case acquisition module 802 is configured to acquire user behavior cases in a pre-set virtual community; the user behavior cases are represented by user behavior attributes and user behavior types;

[0133] A case library construction module 804 is configured to cluster the user behavior cases based on information entropy increase, and construct a source case library according to the clustering results of the user behavior cases;

[0134] A retrieval module 806 is configured to obtain a solution to a target case to be predicted, and retrieve a source case library based on the target case, the case library, and a preset case recommendation threshold;

[0135] A case processing module 808 is configured to reuse a case in the retrieved case set when the retrieved case set is not empty, and revise a source case when the retrieved case set is empty, and determine whether to retain the revised source case based on a similarity between the revised case and the most similar case in the case library;

[0136] The behavior prediction module 810 is used to predict user behavior based on the reused source cases.

[0137] In one embodiment, the user behavior case acquisition module 802 is further configured to acquire user behavior cases cb= in the virtual community according to a preset virtual community.<u,b> ; where u=<rew,pr,pc,op> , rew represents the interval number of the community user incentive value, pr represents the preference level of the community user behavior, pc represents the preference type of the community user behavior, op represents the relational operator, and b represents the community user behavior type; the community user behavior types include: proposing task behavior, solution behavior, like behavior, and dislike behavior.

[0138] In one embodiment, the case library construction module 804 is further configured to obtain the information entropy of the j-th behavior case of the i-th community user as follows:

[0139]

[0140] Among them, case CB i,j Probability CB i，j represents the number of occurrences of the jth type of behavior of the i-th community user, i∈{1..n}, n is the number of community users in the case library, j∈{0, 1, 2, 3}, representing task behavior, solution behavior, like behavior, and dislike behavior respectively;

[0141] Calculate the information entropy of the j-th behavior case based on the information entropy of the j-th behavior case of the i-th community user for:

[0142]

[0143] According to the information entropy of the j-th behavior case The information entropy of the j-th behavior case of the i-th community user is calculated to obtain the information entropy of the k-th case cluster of the j-th behavior for:

[0144]

[0145] According to the information entropy of the k-th case cluster of the j-th behavior Calculate the average entropy of the k-th case cluster of the j-th behavior for:

[0146]

[0147] According to the information entropy of the j-th behavior case Calculate the entropy increase of the k-th case cluster of the j-th behavior as:

[0148]

[0149] When clustering user behavior cases, obtain the community user behavior data SD and the behavior category set B, calculate the behavior case probability of each community user, and select the community user maxP with the largest probability in behavior b as the cluster center SD maxP Put it into the cluster, select the community user case that is most similar to the community user maxP behavior case in the community user behavior data SD based on the principle of information entropy increase, and add it to the cluster until the average entropy of the cluster no longer increases, and move the cluster center SD maxP Add to the case base CB and delete the cluster cases that have been added to the cluster set from the community user behavior data SD.

[0150] In one embodiment, the retrieval module 806 is further configured to obtain the target case, the case library, and a preset case recommendation threshold;

[0151] The similarity function is constructed as:

[0152]

[0153] Among them, u and v represent cases, u i and v i represents the i-th element of the two cases, λ i Represents the similarity calculation weight of the i-th element;

[0154] According to the similarity function, the similarity between the target case and the retrieval case in the case library is compared, and when the similarity is greater than the case recommendation threshold, the retrieval case will be reused.

[0155] In one embodiment, the case processing module 808 is further configured to construct a revision function as follows:

[0156]

[0157] Where t is the target case, t rew represents the incentive value of the target case t, s represents the revised case after revision, u and v represent the adjacent cases respectively; the source case is revised according to the revision function.

[0158] In one embodiment, the case processing module 808 is further configured to retain the revised case when the similarity between the revised case and the most similar case in the source case library is less than or equal to a threshold, and not retain the revised case when the similarity between the revised case and the most similar case in the source case library is greater than the threshold.

[0159] In one embodiment, the method further includes: an evaluation module for evaluating the predicted user behavior.

[0160] Regarding the specific limitations of the virtual community user behavior prediction device based on case reasoning, please refer to the limitations of the virtual community user behavior prediction method based on case reasoning above, which will not be repeated here. The various modules in the above-mentioned virtual community user behavior prediction device based on case reasoning can be implemented in whole or in part by software, hardware, and a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0161] In one embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as follows: Figure 9As shown. The computer device includes a processor, memory, network interface, display screen and input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, a method for predicting user behavior in a virtual community based on case reasoning is implemented. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covering the display screen, or a key, trackball or touchpad provided on the computer device housing, or an external keyboard, touchpad or mouse.

[0162] Those skilled in the art will understand that Figure 9 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0163] In one embodiment, a computer device is provided, including a memory and a processor. The memory stores a computer program, and the processor implements the steps of the method in the above embodiment when executing the computer program.

[0164] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps of the method in the above embodiment are implemented.

[0165] Those skilled in the art will appreciate that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0166] The technical features of the above embodiments can be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0167] The above-described embodiments merely represent several implementation methods of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that a person skilled in the art could make various modifications and improvements without departing from the spirit of the present application, all of which fall within the scope of protection of the present application. Therefore, the scope of protection of the present patent application shall be determined by the appended claims.

Claims

1. A method for predicting user behavior in a virtual community based on case-based reasoning, characterized in that: The method comprises: According to a pre-set virtual community, obtaining user behavior cases in the virtual community; the user behavior cases are represented by user behavior attributes and user behavior types; Clustering the user behavior cases based on information entropy increase, and building a source case library according to the clustering results of the user behavior cases; Obtain a solution to a target case to be predicted, and search the source case library according to the target case, the source case library, and a preset case recommendation threshold; When the retrieved case set is not empty, the cases in the retrieved case set are reused; when the retrieved case set is empty, the source case is revised, and based on the similarity between the revised case and the most similar case in the case library, it is determined whether to retain the revised case; Predict user behavior based on reused source cases.

2. The method according to claim 1, characterized in that The step of obtaining user behavior cases in a pre-set virtual community includes: According to the preset virtual community, obtain the user behavior case cb=<u,b> ; Where u=<rew,pr,pc,op> , rew represents the interval number of the community user incentive value, pr represents the preference level of the community user behavior, pc represents the preference type of the community user behavior, op represents the relational operator, and b represents the community user behavior type; the community user behavior types include: proposing task behavior, solution behavior, like behavior, and dislike behavior.

3. The method according to claim 1, characterized in that Clustering the user behavior cases based on information entropy increase includes: The information entropy of the j-th behavior case of the i-th community user is obtained as: Among them, case CB i,j Probability CB i，j represents the number of occurrences of the jth type of behavior of the i-th community user, i∈{1..n}, n is the number of community users in the case library, j∈{0, 1, 2, 3}, representing task behavior, solution behavior, like behavior, and dislike behavior respectively; Calculate the information entropy of the j-th behavior case based on the information entropy of the j-th behavior case of the i-th community user for: According to the information entropy of the j-th behavior case The information entropy of the j-th behavior case of the i-th community user is calculated to obtain the information entropy of the k-th case cluster of the j-th behavior for: According to the information entropy of the k-th case cluster of the j-th behavior Calculate the average entropy of the k-th case cluster of the j-th behavior for: According to the information entropy of the j-th behavior case Calculate the entropy increase of the k-th case cluster of the j-th behavior as: When clustering user behavior cases, obtain the community user behavior data SD and the behavior category set B, calculate the behavior case probability of each community user, and select the community user maxP with the largest probability in behavior b as the cluster center SD maxP Put it into the cluster, select the community user case that is most similar to the community user maxP behavior case in the community user behavior data SD based on the principle of information entropy increase, and add it to the cluster until the average entropy of the cluster no longer increases, and move the cluster center SD maxP Add to the case base CB and delete the cluster cases that have been added to the cluster set from the community user behavior data SD.

4. The method according to any one of claims 1 to 3, characterized in that Searching the source case library according to the target case, the source case library, and a preset case recommendation threshold includes: Obtaining the target case, the source case library, and a preset case recommendation threshold; The similarity function is constructed as: Among them, u and v represent cases, u i and v i represents the i-th element of the two cases, λ i Represents the similarity calculation weight of the i-th element; According to a similarity function, the similarity between the target case and the retrieved case in the source case library is compared. When the similarity is greater than a case recommendation threshold, the retrieved case is recommended as a solution to the target case.

5. The method according to any one of claims 1 to 3, characterized in that When the search case set is empty, the source case is revised, including: The build revision function is: Where t is the target case, t rew represents the incentive value of the target case t, s represents the revised case after revision, u and v represent the adjacent source cases respectively; The source case is revised according to the revision function.

6. The method according to claim 5, characterized in that Determine whether to retain the revised case based on its similarity to the most similar case in the case database, including: When the similarity between the revised case and the most similar case in the source case library is less than or equal to the threshold, the revised case is retained. When the similarity between the revised case and the most similar case in the source case library is greater than the threshold, the revised case is not retained.

7. The method according to claim 6, characterized in that The method further comprises: Evaluate predicted user behavior.

8. A virtual community user behavior prediction device based on case-based reasoning, characterized in that: The device comprises: A user behavior case acquisition module is used to acquire user behavior cases in a pre-set virtual community according to the virtual community; the user behavior cases are represented by user behavior attributes and user behavior types; A case library construction module is used to cluster the user behavior cases based on information entropy increase, and to construct a source case library according to the clustering results of the user behavior cases; A retrieval module is used to obtain a solution to a target case to be predicted, and retrieve the source case library according to the target case, the source case library, and a preset case recommendation threshold; a case processing module, configured to reuse a case in the retrieved case set when the retrieved case set is not empty, and revise a source case when the retrieved case set is empty, and determine whether to retain the revised source case based on a similarity between the revised case and the most similar case in the case library; The behavior prediction module is used to predict user behavior based on reused source cases.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images