Virtual resource recommendation method and device, equipment, storage medium and program product

Abstract

The invention relates to a virtual resource recommendation method and device, equipment, a storage medium and a program product. The method comprises the following steps: acquiring user information of a user, historical transaction information and resource characteristics of a plurality of virtual resources, and determining a plurality of decision results for representing the transaction probability of the user according to the user information, the historical transaction information and the plurality of resource characteristics by using a multi-level decision tree model, and determining a target virtual resource to be recommended from the plurality of virtual resources by utilizing a long-short-term memory network model according to the plurality of decision results. By adopting the method, the accuracy of the recommended target virtual resources can be improved, and the virtual resources suitable for the users can be recommended for different users.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and in particular to a virtual resource recommendation method, device, equipment, storage medium and program product. Background Art

[0002] At present, the scale of virtual resources (such as fund products) occupies an important position in the resource market. Users can choose the virtual resources according to their own risk tolerance and goals.

[0003] Since there are many types of virtual resources, it is difficult for users to choose. Therefore, how to recommend virtual resources suitable for different users has become an urgent problem to be solved by people in this field. Summary of the invention

[0004] Based on this, it is necessary to provide a virtual resource recommendation method, device, equipment, storage medium and program product that can recommend virtual resources suitable for different users to address the above technical problems.

[0005] In a first aspect, the present application provides a virtual resource recommendation method. The method comprises:

[0006] Obtaining user information, historical transaction information, and resource characteristics of multiple virtual resources;

[0007] Determine multiple decision results based on the user information, the historical transaction information and multiple resource features using a multi-level decision tree model; the decision results are used to characterize the transaction probability of the user;

[0008] A long short-term memory network model is used to determine a target virtual resource to be recommended from the multiple virtual resources according to the multiple decision results.

[0009] In one embodiment, the multi-level decision tree model is used to determine multiple decision results according to the user information, the historical transaction information and multiple resource characteristics, including:

[0010] Inputting the user information, the historical transaction information and the plurality of resource features into a first-level decision tree model of the multi-level decision tree model to obtain a decision result output by the first-level decision tree model;

[0011] In addition to the first-level decision tree model, the decision result output by the previous-level decision tree model and the user information, the historical transaction information and a plurality of resource features are input into the next-level decision tree model to obtain the decision result output by the next-level decision tree model.

[0012] In one embodiment, the decision result output by the previous decision tree model and the user information, the historical transaction information and the plurality of resource features are input into the next decision tree model to obtain the decision result output by the next decision tree model, including:

[0013] The decision result output by the previous decision tree model is concatenated with the user information, the historical transaction information and a plurality of resource features to obtain a vector matrix; wherein different row vectors of the vector matrix correspond to different information;

[0014] The vector matrix is ​​input into the next-level decision tree model to obtain the decision result output by the next-level decision tree model.

[0015] In one embodiment, the long short-term memory network model is used to determine the target virtual resource to be recommended from the multiple virtual resources according to the multiple decision results, including:

[0016] Inputting a plurality of the decision results into the long short-term memory network model to obtain a prediction result; the prediction result is used to characterize the predicted transaction probability of the virtual resource;

[0017] The target recommended resource is determined from a plurality of virtual resources according to the prediction result and the preset recommendation condition.

[0018] In one embodiment, the preset recommendation condition includes a preset recommendation quantity; and determining the target recommended resource from a plurality of virtual resources according to the prediction result and the preset recommendation condition includes:

[0019] Sorting the plurality of virtual resources according to the prediction result to obtain a sorting result;

[0020] According to the ranking result, a virtual resource matching the preset recommended quantity is selected from the plurality of virtual resources, and the selected virtual resource is determined as the target recommended resource.

[0021] In one embodiment, the preset recommendation condition includes a preset recommendation threshold; and determining the target recommended resource from a plurality of virtual resources according to the prediction result and the preset recommendation condition includes:

[0022] The virtual resource whose prediction result is greater than a preset recommendation threshold is determined as the target recommended resource.

[0023] In one embodiment, the method further comprises:

[0024] Acquire a sample set; the sample set includes sample user information, sample historical transaction information, and sample resource characteristics of multiple sample virtual resources;

[0025] The initial decision tree model and the initial memory network model are trained based on the sample set to obtain the multi-level decision tree model and the long short-term memory network model.

[0026] In a second aspect, the present application also provides a virtual resource recommendation device. The device includes:

[0027] An information acquisition module, used to acquire user information, historical transaction information, and resource characteristics of multiple virtual resources;

[0028] A decision result determination module, used to determine multiple decision results according to the user information, the historical transaction information and multiple resource features using a multi-level decision tree model; the decision results are used to characterize the transaction probability of the user;

[0029] The recommendation module is used to determine a target virtual resource to be recommended from the multiple virtual resources according to the multiple decision results by using a long short-term memory network model.

[0030] In a third aspect, the present application further provides a computer device, which includes a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the above method when executing the computer program.

[0031] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, and the computer program implements the steps of the above method when executed by a processor.

[0032] In a fifth aspect, the present application further provides a computer program product, which includes a computer program, and when the computer program is executed by a processor, the steps of the above method are implemented.

[0033] The above-mentioned virtual resource recommendation method, device, equipment, storage medium and program product obtain the user information, historical transaction information and resource characteristics of multiple virtual resources of the user, use the multi-level decision tree model to determine multiple decision results for characterizing the transaction probability of the user according to the user information, historical transaction information and multiple resource characteristics, and use the long short-term memory network model to determine the target virtual resource to be recommended from multiple virtual resources according to the multiple decision results. At present, the user selects the virtual resources by himself, but the types and quantities of virtual resources are numerous, and it is difficult to select the virtual resources suitable for the user himself. In this application, various information of the user is taken into account, and the trained multi-level decision tree model and the long short-term memory network model are used to determine the target virtual resource to be recommended. Since the combination of the multi-level decision tree model and the long short-term memory network model can extract richer features and obtain more accurate prediction results, it can improve the accuracy of the recommended target virtual resources and recommend virtual resources suitable for different users. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 An application environment diagram of a virtual resource recommendation method provided in an embodiment of the present application;

[0035] Figure 2 One of the flowcharts of a virtual resource recommendation method provided in an embodiment of the present application;

[0036] Figure 3 One of the flowcharts of a method for determining a decision result provided in an embodiment of the present application;

[0037] Figure 4 A second flowchart of a method for determining a decision result provided in an embodiment of the present application;

[0038] Figure 5 One of the flow diagrams of a method for determining a target virtual resource provided in an embodiment of the present application;

[0039] Figure 6 A second flowchart of a method for determining a target virtual resource provided in an embodiment of the present application;

[0040] Figure 7 A second flowchart of a virtual resource recommendation method provided in an embodiment of the present application;

[0041] Figure 8 A third flowchart of a virtual resource recommendation method provided in an embodiment of the present application;

[0042] Fig. 9 A model structure diagram of a virtual resource recommendation method provided in an embodiment of the present application;

[0043] Fig.10 A structural block diagram of a virtual resource recommendation device provided in an embodiment of the present application;

[0044] Fig.11 An internal structure diagram of a computer device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0045] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0046] Figure 1 FIG. 1 is an application environment diagram of a virtual resource recommendation method provided in an embodiment of the present application. The virtual resource recommendation method provided in an embodiment of the present application can be applied to Figure 1The server 101 in the application environment shown can also be used in a system including the server 101 and the terminal 102, which is implemented through the interaction between the server 101 and the terminal 102. The server 101 can be implemented as an independent server or a server cluster composed of multiple servers. The terminal 102 can be, but is not limited to, various personal computers, laptops, smart phones, tablet computers, etc.

[0047] In one embodiment, Figure 2 One of the flow charts of a virtual resource recommendation method provided in an embodiment of the present application is as follows: Figure 2 As shown, a virtual resource recommendation method is provided, which is applied to Figure 1 Taking the server 101 in the example as an example, the following steps may be included:

[0048] S201. Obtain user information, historical transaction information, and resource characteristics of multiple virtual resources.

[0049] The user information may include the user's gender, age, and the risk level of virtual resources that the user can accept for trading. The historical transaction information may include the information corresponding to the historical recommended virtual resources that the user has completed. The resource characteristics may include the risk level of the virtual resource, the theme name of the virtual resource, and the type of the virtual resource.

[0050] In an embodiment of the present application, the user information, historical transaction information and resource characteristics of multiple virtual resources of the user can be sent to the server through the terminal, and the server obtains the user information, historical transaction information and resource characteristics of multiple virtual resources of the user sent by the terminal.

[0051] Alternatively, the server crawls the user information, historical transaction information, and resource characteristics of multiple virtual resources from a website, wherein the website may be a website used by the user of the virtual resource to be recommended.

[0052] S202. Determine multiple decision results based on user information, historical transaction information, and multiple resource features using a multi-level decision tree model; the decision results are used to characterize the transaction probability of the user.

[0053] The multi-level decision tree model is a decision tree model composed of multiple levels, and the multi-level decision tree model can be obtained by training an initial multi-level decision tree model through historical samples.

[0054] In an embodiment of the present application, the server can fill in missing values ​​in historical transaction information. For example, mean interpolation, similar mean interpolation, modeling prediction, high-dimensional mapping, multiple interpolation, etc. can be used to fill in missing values. The server can also encrypt historical transaction information using a preset encryption desensitization algorithm. For example, the preset encryption desensitization algorithm may include a symmetric encryption algorithm, an asymmetric encryption algorithm, a linear hashing algorithm, a hybrid encryption algorithm, etc. The server can input user information, historical transaction information, and multiple resource features into a trained multi-level decision tree model. The multi-level decision tree model can output a decision result that is the same as the level according to the level of the multi-level decision model. Among them, each decision result can be used to represent the transaction probability of the user.

[0055] For example, user information includes gender and the risk level of virtual resources that the user can accept for trading. User information is recorded as X1, gender is recorded as X1-1, X1-1 is 1 when the gender is female, and X1-1 is 0 when the gender is male; the risk level of virtual resources that the user can accept for trading is recorded as X1-2, and the risk level of virtual resources that the user can accept for trading includes 1-5 levels, recorded as 1-5 respectively. Historical transaction information is recorded as X2, X2 is 1 when trading, and X2 is 0 when not trading. Resource characteristics may include the risk level of virtual resources, the theme name of virtual resources, and the type of virtual resources. Resource characteristics are recorded as X3, and the risk level of virtual resources is recorded as X3-1, and the risk level of virtual resources includes 1-5 levels, recorded as 1-5 respectively; the theme name of virtual resources is recorded as X3-2, and the theme names of virtual resources include consumption, new energy, medicine, TMT, cycle, financial real estate, military industry, high-end manufacturing, Hong Kong stocks, etc., which are 1-9 respectively; the type of virtual resources is recorded as X3-3, and the types of virtual resources include value type, growth type, and balance type. Then there are input characteristics X (X1, X2, X3).

[0056] Input the input feature X (X1, X2, X3) into the multi-level decision tree model. If the multi-level decision tree model has 5 levels, the multi-level decision tree model outputs 5 decision results.

[0057] S203: Determine a target virtual resource to be recommended from multiple virtual resources according to multiple decision results using a long short-term memory network model.

[0058] In an embodiment of the present application, the server may input multiple decision results output by the multi-level decision tree model into a long short-term memory network model, and the long short-term memory network (LSTM) model may output the prediction result of the virtual resource. The prediction result may represent the transaction probability of the user trading the virtual resource. The above method may be used to obtain the prediction results of multiple virtual resources, and the server may determine the target virtual resource to be recommended from the multiple virtual resources according to actual needs. The actual needs may include the length of the recommendation list in the recommendation application, the number of recommendations set by the user, etc.

[0059] Continuing with the above example, the server can input the five decision results output by the multi-level decision tree model into the long short-term memory network model, and obtain a prediction result of 30% for virtual resource A, that is, the transaction probability of virtual resource A is 30%. The transaction probabilities of multiple virtual resources can be obtained through the above multi-level decision tree model and the long short-term memory network model. For example, the prediction result of virtual resource A is 30%, the prediction result of virtual resource B is 35%, the prediction result of virtual resource C is 15%, the prediction result of virtual resource D is 68%, the prediction result of virtual resource E is 10%, the prediction result of virtual resource F is 5%, and the prediction result of virtual resource G is 54%. If the length of the recommendation list in the recommendation application or the number of recommendations set by the user is 3, virtual resource B, virtual resource D, and virtual resource G can be determined as the target virtual resources to be recommended.

[0060] In the above virtual resource recommendation method, the user information, historical transaction information and resource characteristics of multiple virtual resources of the user are obtained, and a multi-level decision tree model is used to determine multiple decision results for characterizing the transaction probability of the user according to the user information, historical transaction information and multiple resource characteristics, and a long short-term memory network model is used to determine the target virtual resource to be recommended from multiple virtual resources according to the multiple decision results. At present, the user selects the virtual resources by himself, but the types and quantities of virtual resources are numerous, and it is difficult to select the virtual resources suitable for the user himself. In this application, a variety of information of the user is taken into account, and the trained multi-level decision tree model and the long short-term memory network model are used to determine the target virtual resource to be recommended. Since the combination of the multi-level decision tree model and the long short-term memory network model can extract richer features and obtain more accurate prediction results, it can improve the accuracy of the recommended target virtual resources and recommend virtual resources suitable for different users.

[0061] In one embodiment, Figure 3 One of the flow charts of a method for determining a decision result provided in an embodiment of the present application is as follows: Figure 3As shown, this embodiment relates to a possible implementation method of how to use a multi-level decision tree model to determine multiple decision results according to user information, historical transaction information and multiple resource characteristics. On the basis of the above embodiment, the above S202 includes:

[0062] S301, inputting user information, historical transaction information and multiple resource features into the first-level decision tree model of the multi-level decision tree model to obtain the decision result output by the first-level decision tree model.

[0063] In an embodiment of the present application, the server may preprocess the user information, historical transaction information, and multiple resource features, and input the preprocessed user information, historical transaction information, and multiple resource features into the first-level decision tree model of the multi-level decision tree model, and the first-level decision tree model of the multi-level decision tree model may output a decision result. The preprocessing may include missing value filling, noise processing, etc.

[0064] S302. In addition to the first-level decision tree model, the decision result output by the previous-level decision tree model and user information, historical transaction information, and multiple resource features are input into the next-level decision tree model to obtain the decision result output by the next-level decision tree model.

[0065] In the embodiment of the present application, in addition to the first-level decision tree model, the server can splice the decision result output by the previous-level decision tree model with user information, historical transaction information, and multiple resource features to obtain a splicing vector. The server inputs the splicing vector into the next-level decision tree model to obtain the decision result output by the next-level decision tree model.

[0066] For example, user information includes gender and the risk level of virtual resources that the user can accept for trading. User information is recorded as X1, gender is recorded as X1-1, X1-1 is 1 when the gender is female, and X1-1 is 0 when the gender is male; the risk level of virtual resources that the user can accept for trading is recorded as X1-2, and the risk level of virtual resources that the user can accept for trading includes 1-5 levels, recorded as 1-5 respectively. Historical transaction information is recorded as X2, X2 is 1 when trading, and X2 is 0 when not trading. Resource characteristics may include the risk level of virtual resources, the theme name of virtual resources, and the type of virtual resources. Resource characteristics are recorded as X3, and the risk level of virtual resources is recorded as X3-1, and the risk level of virtual resources includes 1-5 levels, recorded as 1-5 respectively; the theme name of virtual resources is recorded as X3-2, and the theme names of virtual resources include consumption, new energy, medicine, TMT, cycle, financial real estate, military industry, high-end manufacturing, Hong Kong stocks, etc., which are 1-9 respectively; the type of virtual resources is recorded as X3-3, and the types of virtual resources include value type, growth type, and balance type. The decision result output by the previous level decision tree model is recorded as F.

[0067] Therefore, the decision result output by the previous decision tree model is spliced ​​with user information, historical transaction information and multiple resource features, and the spliced ​​vector obtained after splicing is W (F, X1-1, X1-2, X2, X3-1, X3-2, X3-3). W (F, X1-1, X1-2, X2, X3-1, X3-2, X3-3) is input into the next decision tree model to obtain the decision result output by the next decision tree model.

[0068] Specifically, when the multi-level decision model has 3 levels, the server first inputs W1 (X1-1, X1-2, X2, X3-1, X3-2, X3-3) into the first-level decision tree model to obtain F1; then concatenates F1 with X1-1, X1-2, X2, X3-1, X3-2, X3-3 to obtain W2 (F1, X1-1, X1-2, X2, X3-1, X3-2, X3-3), and inputs W2 into the second-level decision tree model to obtain F2; then concatenates F2 with F1, X1-1, X1-2, X2, X3-1, X3-2, X3-3 to obtain W3 (F1, F2, X1-1, X1-2, X2, X3-1, X3-2, X3-3), and inputs W3 into the second-level decision tree model to obtain F3. The decision results obtained by the multi-level decision model include F1, F2 and F3.

[0069] In this embodiment, the server inputs user information, historical transaction information, and multiple resource features into the first-level decision tree model of the multi-level decision tree model to obtain the decision result output by the first-level decision tree model. In addition to the first-level decision tree model, the decision result output by the previous-level decision tree model and the user information, historical transaction information, and multiple resource features are input into the next-level decision tree model to obtain the decision result output by the next-level decision tree model. Since the information input into the model has high dimensionality, the first-level decision tree model can only learn fewer features, while the multi-level decision tree model can learn more feature-rich features. Therefore, inputting user information, historical transaction information, and multiple resource features into the multi-level decision tree model can improve the accuracy of the decision results.

[0070] In one embodiment, Figure 4 A second flow chart of a method for determining a decision result provided in an embodiment of the present application is as follows: Figure 4 As shown, this embodiment involves a possible implementation method of how to input the decision result output by the previous level decision tree model and user information, historical transaction information and multiple resource features into the next level decision tree model to obtain the decision result output by the next level decision tree model. On the basis of the above embodiment, the above S302 includes:

[0071] S401. Concatenate the decision result output by the previous level decision tree model with the user information, historical transaction information and multiple resource features to obtain a vector matrix; wherein different row vectors of the vector matrix correspond to different information.

[0072] In an embodiment of the present application, the server can splice the decision results output by the previous level decision tree model with user information, historical transaction information and multiple resource characteristics. After splicing, a vector matrix can be obtained. Different row vectors of the vector matrix can respectively correspond to the decision results, user information, historical transaction information and resource characteristics output by the previous level decision tree model.

[0073] Following the example of user information, historical transaction information, and resource feature content in S302, the server can splice the decision result output by the previous level decision tree model with the user information, historical transaction information, and multiple resource features. The vector matrix V1 obtained after splicing is:

[0074]

[0075] The row features in the vector matrix include decision result row features, user information row features, historical transaction row features, and resource row features. The decision result row features include multiple decision result vectors, the user information row features include multiple user information vectors, the historical transaction row features include multiple historical transaction vectors, and the resource row features include multiple resource vectors.

[0076] S402: Input the vector matrix into the next-level decision tree model to obtain the decision result output by the next-level decision tree model.

[0077] In an embodiment of the present application, the server may input the vector matrix into a subsequent decision tree model, and the subsequent decision tree model outputs a decision result corresponding to the subsequent decision tree model.

[0078] For example, if the multi-level decision tree model is 3 levels, the decision result output by the first-level decision tree model is concatenated with user information, historical transaction information, and multiple resource features to obtain a vector matrix V1. Then, the vector matrix V1 is input into the second-level decision tree model to obtain the decision result F2 output by the second-level decision tree model. Then, F2 is concatenated with the vector matrix V1 to obtain the vector matrix V2:

[0079]

[0080] Then the vector matrix V2 is input into the third-level decision tree model to obtain the decision result F3 output by the third-level decision tree model. Then the decision results obtained by the multi-level decision model include F1, F2 and F3.

[0081] In this embodiment, the decision results output by the previous level decision tree model are spliced ​​with user information, historical transaction information and multiple resource features to obtain a vector matrix; wherein different row vectors of the vector matrix correspond to different information, and the vector matrix is ​​input into the next level decision tree model to obtain the decision results output by the next level decision tree model, which can improve the accuracy of the decision results.

[0082] In one embodiment, Figure 5 One of the flow charts of a method for determining a target virtual resource provided in an embodiment of the present application is as follows: Figure 5 As shown, this embodiment relates to a possible implementation method of how to use a long short-term memory network model to determine a target virtual resource to be recommended from multiple virtual resources according to multiple decision results. On the basis of the above embodiment, the above S203 includes:

[0083] S501. Input multiple decision results into a long short-term memory network model to obtain a prediction result; the prediction result is used to characterize the predicted transaction probability of the virtual resource.

[0084] Among them, the long short-term memory network model is a neural network model containing long short-term memory blocks.

[0085] In an embodiment of the present application, the server can input a decision result into a long short-term memory network model, and the long short-term memory network model can output a prediction result, and the output prediction result can represent the predicted transaction probability of the virtual resource.

[0086] For example, if the pole number of the multi-level decision tree model is 3, then the multi-level decision tree model is used to obtain 5 decision results for virtual resource A. The 5 decision results are input into the long short-term memory network model, and the long short-term memory network model outputs 1 prediction result of 64.5%, that is, the predicted transaction probability of virtual resource A is 64.5%.

[0087] S502: Determine a target recommended resource from a plurality of virtual resources according to the prediction result and preset recommendation conditions.

[0088] In the embodiment of the present application, the server may screen multiple virtual resources according to the prediction results of the multiple virtual resources and preset recommendation conditions to determine the target recommended resource. The preset recommendation conditions may include a preset recommendation quantity, a preset recommendation threshold, etc.

[0089] Case 1: When the preset recommendation conditions include the preset recommendation quantity, such as Figure 6 As shown, Figure 6 The second flowchart of a method for determining a target virtual resource provided in an embodiment of the present application may include the following steps:

[0090] S601. Sort multiple virtual resources according to prediction results to obtain sorting results.

[0091] In the embodiment of the present application, the server can sort multiple virtual resources according to the size of the numerical value in the prediction result to obtain a sorting result. For example, the prediction result of virtual resource A is 25%, the prediction result of virtual resource B is 13%, the prediction result of virtual resource C is 72%, the prediction result of virtual resource D is 18%, and the prediction result of virtual resource E is 51%, then the sorting result is: virtual resource C, virtual resource E, virtual resource A, virtual resource D, virtual resource B.

[0092] S602: According to the sorting result, select virtual resources matching a preset recommended quantity from multiple virtual resources, and determine the selected virtual resources as target recommended resources.

[0093] In an embodiment of the present application, the server can select virtual resources that match the preset recommended quantity from multiple virtual resources according to the position of each virtual resource in the sorting result, and determine the selected virtual resources as target recommended resources. For example, if the sorting result is: virtual resource C, virtual resource E, virtual resource A, virtual resource D, virtual resource B, and the preset recommended quantity is 2, then virtual resource C and virtual resource E can be selected from virtual resource A, virtual resource B, virtual resource C, virtual resource D, and virtual resource E as target recommended resources.

[0094] Case 2: When the preset recommendation condition includes a preset recommendation threshold, a virtual resource whose prediction result is greater than the preset recommendation threshold is determined as a target recommended resource.

[0095] Specifically, when the preset recommendation condition includes a preset recommendation threshold, the server can select a virtual resource whose prediction result is greater than the preset recommendation threshold from multiple virtual resources, and use the virtual resource whose prediction result is greater than the preset recommendation threshold as the target recommendation resource. For example, if the prediction result of virtual resource A is 25%, the prediction result of virtual resource B is 13%, the prediction result of virtual resource C is 72%, the prediction result of virtual resource D is 18%, and the prediction result of virtual resource E is 51%, and the preset recommendation threshold is 50%, virtual resource C and virtual resource E can be used as target recommendation resources.

[0096] In the above embodiment, multiple decision results are input into the long short-term memory network model to obtain a predicted transaction probability prediction result for characterizing virtual resources, and a target recommended resource is determined from multiple virtual resources based on the prediction result and the preset recommendation condition. Wherein, when the preset recommendation condition includes a preset recommendation quantity, a virtual resource matching the preset recommendation quantity is selected from multiple virtual resources based on the sorting result, and the selected virtual resource is determined as the target recommended resource; when the preset recommendation condition includes a preset recommendation threshold, a virtual resource whose prediction result is greater than the preset recommendation threshold is determined as the target recommended resource. Since the preset recommendation condition can be set according to actual needs, in the embodiment of the present application, the target recommended resource can be screened from multiple virtual resources through the preset recommendation condition, so that a virtual resource that is more suitable for the user himself can be found.

[0097] In one embodiment, Figure 7 A second flowchart of a virtual resource recommendation method provided in an embodiment of the present application is shown in FIG. Figure 7 As shown, based on the above embodiment, the method further includes:

[0098] S701, obtaining a sample set; the sample set includes sample user information of the user, sample historical transaction information, and sample resource characteristics of multiple sample virtual resources.

[0099] In an embodiment of the present application, the terminal may send a sample set including the user's sample user information, sample historical transaction information, and sample resource characteristics of multiple sample virtual resources to the server, and the server may obtain the sample set in this way. The server may also directly obtain the user's sample user information, sample historical transaction information, and sample resource characteristics of multiple sample virtual resources stored in the database.

[0100] S702: Train the initial decision tree model and the initial memory network model based on the sample set to obtain a multi-level decision tree model and a long short-term memory network model.

[0101] In an embodiment of the present application, the server can input the sample data in the sample set into the initial prediction model which is a combination of the initial decision tree model and the initial memory network model to obtain a first training result. The first training result is compared with the actual result corresponding to the sample data. If the loss value between the first prediction result and the actual result is greater than the preset result threshold, the parameters in the initial prediction model are adjusted, and another sample data in the sample set is input into the adjusted prediction model to obtain a second prediction result. Referring to the above, if the loss value between the second prediction result and the actual result is greater than the preset result threshold, the parameters in the prediction model continue to be adjusted; if the loss value between the second prediction result and the actual result is not greater than the preset result threshold, the training is terminated, and the prediction model at the end is determined as the target prediction model, which includes a multi-level decision tree model and a long short-term memory network model.

[0102] Multiple verification data can be input into the target prediction model to obtain multiple verification results; the target prediction model can be evaluated based on the multiple verification results to adjust the model parameters or model structure according to the evaluation results. The target prediction model can be evaluated based on preset indicators and multiple verification results. The preset indicators can include the accuracy of the model, the model calculation speed, etc. In addition, if the accuracy of the model is less than a preset accuracy threshold, or the model calculation speed is less than a preset speed threshold, the model parameters or model structure can be adjusted.

[0103] For example, the verification results may include true positive (TP), false positive (FP—FalsePositive), true negative (TN—TrueNegative), false negative (FN—FalseNegative), etc. Among them, TP means predicted positive and actually positive. The positive sample is successfully predicted as positive. FP means predicted positive and actually negative. The negative sample is incorrectly predicted as positive. TN means predicted negative and actually negative. The negative sample is successfully predicted as negative. FN means predicted negative and actually positive. The positive sample is incorrectly predicted as negative. The accuracy of the model can be calculated using the following formula.

[0104] accurac =(TP+TN) / (TP+FP+TN+FN)

[0105] Assuming that the preset accuracy threshold is 98%, if the accuracy of the model is less than 98%, the model structure is adjusted. For example, the number of levels in the initial decision tree model can be increased. In addition, if the model calculation speed is less than the preset speed threshold, the number of levels in the initial decision tree model can be reduced.

[0106] It should be noted that if the difference between the accuracy of the verification result of the initial decision tree model before increasing the number of levels and the accuracy of the verification result of the initial decision tree model after increasing the number of levels is less than a preset threshold, the initial decision tree model before increasing the number of levels can be determined as the initial decision tree model of the trained multi-level decision tree model. Since the model architecture of the initial decision tree model with a small number of levels is small, the model training speed and prediction speed are fast, so the efficiency of virtual resource recommendation can be improved.

[0107] In a detailed embodiment, Figure 8 The third flowchart of a virtual resource recommendation method provided in an embodiment of the present application is as follows: Fig. 9 A model structure diagram of a virtual resource recommendation method provided in an embodiment of the present application, the method comprising:

[0108] S801, obtaining a sample set; the sample set includes sample user information of the user, sample historical transaction information, and sample resource characteristics of multiple sample virtual resources.

[0109] S802: Train the initial decision tree model and the initial memory network model based on the sample set to obtain a multi-level decision tree model and a long short-term memory network model.

[0110] S803: Obtain user information, historical transaction information, and resource characteristics of multiple virtual resources.

[0111] S804: Input user information, historical transaction information, and multiple resource features into a first-level decision tree model of a multi-level decision tree model to obtain a decision result output by the first-level decision tree model.

[0112] S805. In addition to the first-level decision tree model, the decision results output by the previous-level decision tree model are concatenated with user information, historical transaction information, and multiple resource features to obtain a vector matrix; wherein different row vectors of the vector matrix correspond to different information.

[0113] S806: Input the vector matrix into the next-level decision tree model to obtain the decision result output by the next-level decision tree model.

[0114] S807, input multiple decision results into the long short-term memory network model to obtain a prediction result; the prediction result is used to characterize the predicted transaction probability of the virtual resource. If the preset recommendation condition includes a preset recommendation quantity, S808 is executed; if the preset recommendation condition includes a preset recommendation threshold, S809 is executed.

[0115] S808. Sort the multiple virtual resources according to the prediction result to obtain a sorting result. According to the sorting result, select virtual resources matching the preset recommended quantity from the multiple virtual resources, and determine the selected virtual resources as target recommended resources.

[0116] S809: Determine the virtual resource whose prediction result is greater than a preset recommendation threshold as a target recommended resource.

[0117] In an embodiment of the present application, a sample set is obtained; the sample set includes sample user information of the user, sample historical transaction information, and sample resource features of multiple sample virtual resources. The initial decision tree model and the initial memory network model are trained based on the sample set to obtain a multi-level decision tree model and a long short-term memory network model. The user information, historical transaction information, and resource features of multiple virtual resources of the user are obtained. The user information, historical transaction information, and multiple resource features are input into the first-level decision tree model of the multi-level decision tree model to obtain the decision result output by the first-level decision tree model. In addition to the first-level decision tree model, the decision result output by the previous-level decision tree model is spliced ​​with the user information, historical transaction information, and multiple resource features to obtain a vector matrix; wherein different row vectors of the vector matrix correspond to different information. The vector matrix is ​​input into the next-level decision tree model to obtain the decision result output by the next-level decision tree model. Multiple decision results are input into the long short-term memory network model to obtain a prediction result. Among them, multiple virtual resources can be sorted according to the prediction results to obtain the sorting results. According to the sorting results, virtual resources that match the preset recommendation quantity are selected from the multiple virtual resources, and the selected virtual resources are determined as target recommended resources. Virtual resources whose prediction results are greater than the preset recommendation threshold can also be determined as target recommended resources. In this application, various information of users is taken into account, and the trained multi-level decision tree model and long short-term memory network model are used to determine the target virtual resources to be recommended. Since the combination of the multi-level decision tree model and the long short-term memory network model can extract richer features and obtain more accurate prediction results, it is possible to improve the accuracy of the recommended target virtual resources and recommend virtual resources suitable for different users.

[0118] It should be understood that, although the steps in the flowcharts involved in the above embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps is not strictly limited in order, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above embodiments may include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0119] Based on the same inventive concept, the embodiment of the present application also provides a virtual resource recommendation device for implementing the virtual resource recommendation method involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more virtual resource recommendation device embodiments provided below can refer to the limitations of the virtual resource recommendation method above, and will not be repeated here.

[0120] In one embodiment, Fig.10 A structural block diagram of a virtual resource recommendation device provided in an embodiment of the present application, such as Fig.10 As shown, a virtual resource recommendation device 1000 is provided, comprising: an information acquisition module 1001, a decision result determination module 1002 and a recommendation module 1003, wherein:

[0121] The information acquisition module 1001 is used to acquire the user information, historical transaction information and resource characteristics of multiple virtual resources.

[0122] The decision result determination module 1002 is used to determine multiple decision results based on user information, historical transaction information and multiple resource characteristics using a multi-level decision tree model; the decision results are used to characterize the transaction probability of the user.

[0123] The recommendation module 1003 is used to determine the target virtual resource to be recommended from multiple virtual resources according to multiple decision results by using the long short-term memory network model.

[0124] In one embodiment, the decision result determination module 1002 includes:

[0125] The first determination submodule is used to input user information, historical transaction information and multiple resource features into the first-level decision tree model of the multi-level decision tree model to obtain the decision result output by the first-level decision tree model.

[0126] The second determination submodule is used to input the decision results output by the previous level decision tree model, user information, historical transaction information and multiple resource features into the next level decision tree model in addition to the first level decision tree model to obtain the decision results output by the next level decision tree model.

[0127] In one embodiment, the second determination submodule is specifically used to concatenate the decision result output by the previous level decision tree model with user information, historical transaction information and multiple resource features to obtain a vector matrix; wherein different row vectors of the vector matrix correspond to different information; and the vector matrix is ​​input into the next level decision tree model to obtain the decision result output by the next level decision tree model.

[0128] In one embodiment, the recommendation module 1003 includes:

[0129] The third determination submodule is used to input multiple decision results into the long short-term memory network model to obtain a prediction result; the prediction result is used to characterize the predicted transaction probability of the virtual resource.

[0130] The fourth determination submodule is used to determine a target recommended resource from a plurality of virtual resources according to the prediction result and the preset recommendation condition.

[0131] In one embodiment, the preset recommendation condition includes a preset recommendation quantity, and the fourth determination submodule includes:

[0132] The first determining unit is used to sort the multiple virtual resources according to the prediction result to obtain a sorting result.

[0133] The second determining unit is used to select virtual resources matching a preset recommended quantity from a plurality of virtual resources according to the sorting result, and determine the selected virtual resources as target recommended resources.

[0134] In one embodiment, the preset recommendation condition includes a preset recommendation threshold, and the fourth determination submodule includes:

[0135] The third determining unit is configured to determine the virtual resource whose prediction result is greater than a preset recommendation threshold as a target recommended resource.

[0136] In one embodiment, the virtual resource recommendation device 1000 further includes:

[0137] The sample acquisition module is used to acquire a sample set; the sample set includes sample user information, sample historical transaction information, and sample resource characteristics of multiple sample virtual resources.

[0138] The training module is used to train the initial decision tree model and the initial memory network model based on the sample set to obtain a multi-level decision tree model and a long short-term memory network model.

[0139] Each module in the above virtual resource recommendation device can be implemented in whole or in part by software, hardware or a combination thereof. Each module can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute the operations corresponding to each module.

[0140] In one embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as follows: Fig.11 As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. The processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. 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, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store sample data. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a virtual resource recommendation method is implemented.

[0141] Those skilled in the art will understand that Fig.11 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 those shown in the figure, or combine certain components, or have a different arrangement of components.

[0142] In one embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, the following steps are implemented:

[0143] Obtaining user information, historical transaction information, and resource characteristics of multiple virtual resources;

[0144] A multi-level decision tree model is used to determine multiple decision results based on user information, historical transaction information, and multiple resource characteristics; the decision results are used to characterize the transaction probability of the user;

[0145] A long short-term memory network model is used to determine the target virtual resource to be recommended from multiple virtual resources according to multiple decision results.

[0146] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0147] Inputting user information, historical transaction information, and multiple resource features into a first-level decision tree model of a multi-level decision tree model to obtain a decision result output by the first-level decision tree model;

[0148] In addition to the first-level decision tree model, the decision results output by the previous-level decision tree model and user information, historical transaction information, and multiple resource features are input into the next-level decision tree model to obtain the decision results output by the next-level decision tree model.

[0149] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0150] The decision results output by the previous decision tree model are concatenated with user information, historical transaction information, and multiple resource features to obtain a vector matrix; wherein different row vectors of the vector matrix correspond to different information;

[0151] The vector matrix is ​​input into the next-level decision tree model to obtain the decision result output by the next-level decision tree model.

[0152] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0153] Input multiple decision results into the long short-term memory network model to obtain prediction results; the prediction results are used to characterize the predicted transaction probability of virtual resources;

[0154] According to the prediction results and preset recommendation conditions, a target recommended resource is determined from multiple virtual resources.

[0155] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0156] Sorting multiple virtual resources according to the prediction results to obtain a sorting result;

[0157] According to the sorting result, a virtual resource matching the preset recommended quantity is selected from the multiple virtual resources, and the selected virtual resource is determined as the target recommended resource.

[0158] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0159] The virtual resources whose prediction results are greater than a preset recommendation threshold are determined as target recommended resources.

[0160] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0161] Acquire a sample set; the sample set includes sample user information, sample historical transaction information, and sample resource characteristics of multiple sample virtual resources;

[0162] The initial decision tree model and the initial memory network model are trained based on the sample set to obtain a multi-level decision tree model and a long short-term memory network model.

[0163] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented:

[0164] Obtaining user information, historical transaction information, and resource characteristics of multiple virtual resources;

[0165] A multi-level decision tree model is used to determine multiple decision results based on user information, historical transaction information, and multiple resource characteristics; the decision results are used to characterize the transaction probability of the user;

[0166] A long short-term memory network model is used to determine the target virtual resource to be recommended from multiple virtual resources according to multiple decision results.

[0167] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0168] Inputting user information, historical transaction information, and multiple resource features into a first-level decision tree model of a multi-level decision tree model to obtain a decision result output by the first-level decision tree model;

[0169] In addition to the first-level decision tree model, the decision results output by the previous-level decision tree model and user information, historical transaction information, and multiple resource features are input into the next-level decision tree model to obtain the decision results output by the next-level decision tree model.

[0170] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0171] The decision results output by the previous decision tree model are concatenated with user information, historical transaction information, and multiple resource features to obtain a vector matrix; wherein different row vectors of the vector matrix correspond to different information;

[0172] The vector matrix is ​​input into the next-level decision tree model to obtain the decision result output by the next-level decision tree model.

[0173] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0174] Input multiple decision results into the long short-term memory network model to obtain prediction results; the prediction results are used to characterize the predicted transaction probability of virtual resources;

[0175] According to the prediction results and preset recommendation conditions, a target recommended resource is determined from multiple virtual resources.

[0176] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0177] Sorting multiple virtual resources according to the prediction results to obtain a sorting result;

[0178] According to the sorting result, a virtual resource matching the preset recommended quantity is selected from the multiple virtual resources, and the selected virtual resource is determined as the target recommended resource.

[0179] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0180] The virtual resources whose prediction results are greater than a preset recommendation threshold are determined as target recommended resources.

[0181] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0182] Acquire a sample set; the sample set includes sample user information, sample historical transaction information, and sample resource characteristics of multiple sample virtual resources;

[0183] The initial decision tree model and the initial memory network model are trained based on the sample set to obtain a multi-level decision tree model and a long short-term memory network model.

[0184] In one embodiment, a computer program product is provided, comprising a computer program, which, when executed by a processor, implements the following steps:

[0185] Obtaining user information, historical transaction information, and resource characteristics of multiple virtual resources;

[0186] A multi-level decision tree model is used to determine multiple decision results based on user information, historical transaction information, and multiple resource characteristics; the decision results are used to characterize the transaction probability of the user;

[0187] A long short-term memory network model is used to determine the target virtual resource to be recommended from multiple virtual resources according to multiple decision results.

[0188] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0189] Inputting user information, historical transaction information, and multiple resource features into a first-level decision tree model of a multi-level decision tree model to obtain a decision result output by the first-level decision tree model;

[0190] In addition to the first-level decision tree model, the decision results output by the previous-level decision tree model and user information, historical transaction information, and multiple resource features are input into the next-level decision tree model to obtain the decision results output by the next-level decision tree model.

[0191] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0192] The decision results output by the previous decision tree model are concatenated with user information, historical transaction information, and multiple resource features to obtain a vector matrix; wherein different row vectors of the vector matrix correspond to different information;

[0193] The vector matrix is ​​input into the next-level decision tree model to obtain the decision result output by the next-level decision tree model.

[0194] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0195] Input multiple decision results into the long short-term memory network model to obtain prediction results; the prediction results are used to characterize the predicted transaction probability of virtual resources;

[0196] According to the prediction results and preset recommendation conditions, a target recommended resource is determined from multiple virtual resources.

[0197] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0198] Sorting multiple virtual resources according to the prediction results to obtain a sorting result;

[0199] According to the sorting result, a virtual resource matching the preset recommended quantity is selected from the multiple virtual resources, and the selected virtual resource is determined as the target recommended resource.

[0200] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0201] The virtual resources whose prediction results are greater than a preset recommendation threshold are determined as target recommended resources.

[0202] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0203] Acquire a sample set; the sample set includes sample user information, sample historical transaction information, and sample resource characteristics of multiple sample virtual resources;

[0204] The initial decision tree model and the initial memory network model are trained based on the sample set to obtain a multi-level decision tree model and a long short-term memory network model.

[0205] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant laws, regulations and standards of relevant countries and regions.

[0206] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be completed by instructing the relevant hardware through a computer program, and 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 the memory, database or other medium used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but are not limited to this.

[0207] The technical features of the above embodiments may 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.

[0208] The above embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A virtual resource recommendation method, characterized in that: The method comprises: Obtaining user information, historical transaction information, and resource characteristics of multiple virtual resources; Determine multiple decision results based on the user information, the historical transaction information and the multiple resource characteristics using a multi-level decision tree model; the decision results are used to characterize the transaction probability of the user; A long short-term memory network model is used to determine a target virtual resource to be recommended from the multiple virtual resources according to the multiple decision results.

2. The method according to claim 1, characterized in that The method of using a multi-level decision tree model to determine multiple decision results according to the user information, the historical transaction information and the multiple resource characteristics includes: Inputting the user information, the historical transaction information and the plurality of resource features into a first-level decision tree model of the multi-level decision tree model to obtain a decision result output by the first-level decision tree model; In addition to the first-level decision tree model, the decision result output by the previous-level decision tree model and the user information, the historical transaction information and the plurality of resource features are input into the next-level decision tree model to obtain the decision result output by the next-level decision tree model.

3. The method according to claim 2, characterized in that The step of inputting the decision result output by the previous decision tree model, the user information, the historical transaction information, and the plurality of resource features into the next decision tree model to obtain the decision result output by the next decision tree model includes: The decision result output by the previous decision tree model is concatenated with the user information, the historical transaction information and the plurality of resource features to obtain a vector matrix; wherein different row vectors of the vector matrix correspond to different information; The vector matrix is ​​input into the next-level decision tree model to obtain the decision result output by the next-level decision tree model.

4. The method according to claim 1, characterized in that: The method of using the long short-term memory network model to determine the target virtual resource to be recommended from the multiple virtual resources according to the multiple decision results includes: Inputting a plurality of the decision results into the long short-term memory network model to obtain a prediction result; the prediction result is used to characterize the predicted transaction probability of the virtual resource; The target recommended resource is determined from a plurality of virtual resources according to the prediction result and preset recommendation conditions.

5. The method according to claim 4, characterized in that The preset recommendation condition includes a preset recommendation quantity; and determining the target recommended resource from a plurality of virtual resources according to the prediction result and the preset recommendation condition includes: Sorting the plurality of virtual resources according to the prediction result to obtain a sorting result; According to the ranking result, a virtual resource matching the preset recommended quantity is selected from the plurality of virtual resources, and the selected virtual resource is determined as the target recommended resource.

6. The method according to claim 4, characterized in that The preset recommendation condition includes a preset recommendation threshold; and determining the target recommended resource from a plurality of virtual resources according to the prediction result and the preset recommendation condition includes: The virtual resource whose prediction result is greater than a preset recommendation threshold is determined as the target recommended resource.

7. The method according to any one of claims 1 to 6, characterized in that: The method further comprises: Acquire a sample set; the sample set includes sample user information, sample historical transaction information, and sample resource characteristics of multiple sample virtual resources; The initial decision tree model and the initial memory network model are trained based on the sample set to obtain the multi-level decision tree model and the long short-term memory network model.

8. A virtual resource recommendation device, characterized in that: The device comprises: An information acquisition module, used to acquire user information, historical transaction information, and resource characteristics of multiple virtual resources; A decision result determination module, used to determine multiple decision results according to the user information, the historical transaction information and the multiple resource characteristics by using a multi-level decision tree model; the decision results are used to characterize the transaction probability of the user; The recommendation module is used to determine a target virtual resource to be recommended from the multiple virtual resources according to the multiple decision results by using a long short-term memory network model.

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.

11. A computer program product, comprising a computer program, 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