Server performance evaluation method, performance evaluation device, equipment, medium and product

Abstract

The invention provides a server performance evaluation method, a server performance evaluation device, server performance evaluation equipment, a medium and a product, which can be applied to the technical field of data processing and finance, and the method comprises the following steps: obtaining target transaction information of a target product in an ith time period, the target transaction information is determined based on a transaction knowledge graph constructed by initial transaction information of the target product; the target transaction information is input into the target transaction volume prediction model, the predicted transaction volume of the target product in the (i + 1) th time period is output, and the predicted transaction volume represents the access number of the target product; the target transaction volume prediction model is obtained by training an initial recurrent neural network and an initial self-attention network by using multiple pieces of target training transaction information; according to the predicted transaction volume and equipment parameters of a target server storing the target product, a performance evaluation result is generated, and the performance evaluation result represents whether the target server can bear the access number in the (i + 1) th time period or not.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and more specifically, to a server performance evaluation method, a server performance evaluation device, an electronic device, a computer-readable storage medium, and a computer program product. Background Art

[0002] With the continuous development of Internet technology, most service platforms store different types of products in servers to facilitate users to access or purchase products through mobile terminals such as mobile phones. Therefore, predicting the number of independent visitors corresponding to the products launched by the service platform helps the service platform to reasonably plan the performance and capacity of the server for the product.

[0003] In the process of realizing the concept of this application, it was found that the relevant technology would occupy a large amount of computing resources when predicting the number of product visits, and there was a large difference between the predicted number of visits and the actual number of visits. For the service platform, its reference value was poor, and the server performance might not be enough to meet the huge amount of access data, resulting in server crashes. Summary of the invention

[0004] In view of this, the present application provides a server performance evaluation method, a server performance evaluation device, an electronic device, a computer-readable storage medium, and a computer program product.

[0005] One aspect of the present application provides a server performance evaluation method, comprising:

[0006] In response to the transaction volume prediction instruction, target transaction information of the target product in the i-th time period is obtained, wherein the target transaction information is determined by a transaction knowledge graph constructed based on initial transaction information of the target product, and the transaction information represents multiple initial transaction features of the target product in the transaction market;

[0007] Input the target transaction information into a target transaction volume prediction model, and output the predicted transaction volume of the target product in the i+1th time period, wherein the predicted transaction volume represents the number of visits to the target product, and the target transaction volume prediction model is obtained by training an initial recurrent neural network and an initial self-attention network using multiple target training transaction information;

[0008] A performance evaluation result is generated based on the predicted transaction volume and the device parameters of the target server storing the target product, wherein the performance evaluation result indicates whether the target server can bear the number of visits in the i+1th time period.

[0009] According to an embodiment of the present application, the above target transaction information is obtained in the following manner:

[0010] For any historical time period, obtaining a plurality of the above initial transaction features within the above historical time period;

[0011] Constructing the transaction knowledge graph of the target product according to the multiple initial transaction features;

[0012] The target transaction information is determined based on the multiple transaction knowledge graphs, wherein the target transaction information includes multiple target transaction features corresponding to the target product.

[0013] According to an embodiment of the present application, the target transaction information is determined based on the plurality of transaction knowledge graphs, including:

[0014] For each of the transaction knowledge graphs, determining a plurality of intermediate transaction features related to the transaction volume of the target product from the transaction knowledge graph;

[0015] For each of the intermediate transaction features, calculate the ratio between the number of the intermediate transaction features and the number of the transaction knowledge graphs;

[0016] When the above-mentioned proportion is greater than a preset threshold, the above-mentioned intermediate transaction characteristics are determined as the above-mentioned target transaction information.

[0017] According to an embodiment of the present application, before constructing the above-mentioned transaction knowledge graph, it also includes:

[0018] For each of the initial transaction features, preprocess the initial transaction features to obtain preprocessed initial transaction features, wherein the preprocessing includes removing at least one of missing values, abnormal values, and duplicate values;

[0019] The plurality of preprocessed initial transaction features are normalized to obtain normalized initial transaction features, so as to construct the transaction knowledge graph using the initial transaction features.

[0020] According to an embodiment of the present application, the above target transaction volume prediction model is trained in the following manner:

[0021] In response to the model training instruction, a training sample set is obtained, wherein the training sample set includes a plurality of training transaction information corresponding to different training products in the i-th training time period and a transaction volume label corresponding to each of the training transaction information;

[0022] For each of the training transaction information, input the training transaction information into the initial transaction volume prediction model, and output training prediction information, wherein the training prediction information represents the transaction volume prediction information of the training product in the i+1th training time period;

[0023] Calculate the target loss result based on the training prediction information and the transaction volume label corresponding to the training transaction information;

[0024] The model parameters of the initial transaction volume prediction model are iteratively adjusted according to the target loss result to obtain a trained target transaction volume prediction model.

[0025] According to an embodiment of the present application, the above-mentioned training transaction information includes multiple training transaction features.

[0026] According to an embodiment of the present application, the above training transaction information is input into the initial transaction volume prediction model, and the training prediction information is output, including:

[0027] For each of the training transaction information, multiple of the training transaction features are input into an initial recurrent neural network, and a predicted transaction feature corresponding to the training transaction information is output;

[0028] The initial self-attention network is used to perform self-attention weighted processing on the above-mentioned predicted transaction features to obtain the above-mentioned training prediction information.

[0029] According to an embodiment of the present application, the training transaction information is input into the initial transaction volume prediction model, and the training prediction information is output, which also includes:

[0030] In the case where the initial recurrent neural network processes the training transaction information, obtaining an initial weight parameter of each initial neuron in the initial recurrent neural network;

[0031] For each of the above initial neurons, the above initial weight parameters are processed using a gradient descent algorithm to obtain an initial gradient value;

[0032] The initial weight parameters of the initial neurons are updated based on the initial gradient values ​​to obtain an updated initial recurrent neural network.

[0033] According to an embodiment of the present application, the above-mentioned target transaction volume prediction model includes a target recurrent neural network and a target self-attention network.

[0034] According to an embodiment of the present application, the above-mentioned server performance evaluation method further includes:

[0035] Obtaining a target weight parameter of each target neuron in the target recurrent neural network when the target transaction volume prediction model processes the target transaction information;

[0036] For each of the above target neurons, the above target weight parameters are processed using a gradient descent algorithm to obtain a target gradient value;

[0037] The target weight parameters of the target neurons are updated based on the target gradient values ​​to obtain an updated target recurrent neural network, wherein the updated target transaction volume prediction model includes the updated target recurrent neural network and the target self-attention network.

[0038] According to an embodiment of the present application, the above-mentioned server performance evaluation method further includes:

[0039] Get the actual number of visits in the above i+1th time period;

[0040] Generate an optimized loss result based on the above actual number of visits and the above predicted transaction volume;

[0041] The updated target transaction volume prediction model or the target transaction volume prediction model is optimized according to the optimization loss result to obtain a new target transaction volume prediction model.

[0042] Another aspect of the present application provides a server performance evaluation device, comprising:

[0043] an acquisition module, configured to acquire target transaction information of a target product in an i-th time period in response to a transaction volume prediction instruction, wherein the target transaction information is determined by a transaction knowledge graph constructed based on initial transaction information of the target product, and the transaction information represents a plurality of initial transaction features of the target product in a transaction market;

[0044] A prediction module is used to input the target transaction information into a target transaction volume prediction model, and output the predicted transaction volume of the target product in the i+1th time period, wherein the predicted transaction volume represents the number of visits to the target product, and the target transaction volume prediction model is obtained by training an initial recurrent neural network and an initial self-attention network using multiple target training transaction information;

[0045] The evaluation module is used to generate a performance evaluation result based on the predicted transaction volume and the device parameters of the target server storing the target product, wherein the performance evaluation result represents whether the target server can bear the number of visits in the i+1th time period.

[0046] Another aspect of the present application provides an electronic device, comprising:

[0047] one or more processors;

[0048] a memory for storing one or more programs,

[0049] When the one or more programs are executed by the one or more processors, the one or more processors implement the method as described above.

[0050] Another aspect of the present application provides a computer-readable storage medium storing computer-executable instructions, which are used to implement the method described above when executed.

[0051] Another aspect of the present application provides a computer program product, which includes computer executable instructions, and the instructions are used to implement the method as described above when executed.

[0052] According to the embodiment of the present application, by obtaining the target transaction information of the target product in the i-th time period, the target transaction information is input into the target transaction volume prediction model, the predicted transaction volume of the target product in the i+1-th time period is output, and the performance evaluation result is generated according to the predicted transaction volume and the device parameters of the target server storing the target product. Since the type of target transaction information is determined by the transaction knowledge graph constructed by the initial transaction information in the historical time period, the target transaction information with a greater impact relationship is input into the target transaction volume prediction model, which can reduce the data processing amount of the model, thereby reducing the computing resources used by the model. At the same time, the target transaction volume prediction model constructed based on the initial self-attention network can improve the prediction accuracy of the transaction volume, so that a more accurate server performance evaluation result can be obtained. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] The above and other objects, features and advantages of the present application will become more apparent through the following description of the embodiments of the present application with reference to the accompanying drawings, in which:

[0054] Figure 1 An exemplary system architecture to which a server performance evaluation method according to an embodiment of the present application can be applied is shown;

[0055] Figure 2 A flow chart of a server performance evaluation method according to an embodiment of the present application is shown;

[0056] Figure 3 A schematic diagram of a transaction knowledge graph according to an embodiment of the present application is shown;

[0057] Figure 4 A flowchart of a patrol method according to a target transaction volume prediction model according to an embodiment of the present application is shown;

[0058] Figure 5 A data processing flow chart of an initial recurrent neural network according to an embodiment of the present application is shown;

[0059] Figure 6 A block diagram showing a server performance evaluation device according to an embodiment of the present application; and

[0060] Figure 7A block diagram of an electronic device suitable for implementing the method described above according to an embodiment of the present application is shown. DETAILED DESCRIPTION

[0061] Below, embodiments of the present application will be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of the present application. In the following detailed description, for ease of explanation, many specific details are set forth to provide a comprehensive understanding of the embodiments of the present application. However, it is apparent that one or more embodiments may also be implemented without these specific details. In addition, in the following description, descriptions of known structures and technologies are omitted to avoid unnecessary confusion of the concepts of the present application.

[0062] The terms used herein are only for describing specific embodiments and are not intended to limit the present application. The terms "include", "comprising", etc. used herein indicate the existence of the features, steps, operations and / or components, but do not exclude the existence or addition of one or more other features, steps, operations or components.

[0063] All terms (including technical and scientific terms) used herein have the meanings commonly understood by those skilled in the art unless otherwise defined. It should be noted that the terms used herein should be interpreted as having a meaning consistent with the context of this specification and should not be interpreted in an idealized or overly rigid manner.

[0064] When using expressions such as "at least one of A, B, and C, etc.", they should generally be interpreted according to the meaning of the expression commonly understood by those skilled in the art (for example, "a system having at least one of A, B, and C" should include but is not limited to a system having A alone, B alone, C alone, A and B, A and C, B and C, and / or A, B, C, etc.).

[0065] Currently, most businesses still use manual estimation to predict visit volume, but manual estimation often leads to errors, and the historical transaction data used in the estimation usually has a small time range, which is insufficient to cover all transaction data, resulting in poor accuracy in visit volume prediction.

[0066] Other forecasting methods also have the problem of poor forecasting accuracy when using historical transaction information for estimation due to their model architecture and poor quality of input data.

[0067] In view of this, an embodiment of the present application provides a server performance evaluation method, a performance evaluation device, a device, a medium and a product, which can be applied to the fields of data processing and financial technology. The method includes responding to a transaction volume prediction instruction, obtaining target transaction information of a target product in the i-th time period, wherein the target transaction information is determined based on a transaction knowledge graph constructed based on initial transaction information of the target product, and the transaction information represents a variety of initial transaction features of the target product in the transaction market; inputting the target transaction information into a target transaction volume prediction model, and outputting the predicted transaction volume of the target product in the i+1-th time period, wherein the predicted transaction volume represents the number of visits to the target product, and the target transaction volume prediction model is obtained by training an initial recurrent neural network and an initial self-attention network using multiple target training transaction information; generating a performance evaluation result based on the predicted transaction volume and the device parameters of the target server storing the target product, wherein the performance evaluation result represents whether the target server can carry the number of visits in the i+1-th time period.

[0068] It should be noted that the server performance evaluation method and device provided in the present application can be used in the field of financial technology, such as banks and other financial institutions, and can also be used in any field other than the field of financial technology, such as the field of logistics. Therefore, the application field of the server performance evaluation method and device provided in the present application is not limited.

[0069] In the technical solution of the present application, the user information (including but not limited to user personal information, user image information, user device information, such as location information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved are all information and data authorized by the user or fully authorized by all parties, and the collection, storage, use, processing, transmission, provision, disclosure and application of the relevant data comply with relevant laws, regulations and standards, take necessary confidentiality measures, do not violate public order and good morals, and provide corresponding operation entrances for users to choose to authorize or refuse.

[0070] In the scenario of using personal information for automated decision-making, the methods, devices, and systems provided in the embodiments of the present application provide users with corresponding operation portals for users to choose to agree or reject the results of automated decision-making; if the user chooses to reject, the expert decision-making process will be entered. The expression "automated decision-making" here refers to the activity of automatically analyzing and evaluating an individual's behavioral habits, interests and hobbies, or economic, health, credit status, etc. through a computer program, and making decisions. The expression "expert decision-making" here refers to the activity of making decisions by people who specialize in a certain field, have specialized experience, knowledge and skills, and have reached a certain level of professionalism.

[0071] Figure 1An exemplary system architecture 100 to which a server performance evaluation method according to an embodiment of the present application can be applied is shown. It should be noted that: Figure 1 What is shown is merely an example of a system architecture to which the embodiments of the present application can be applied, in order to help those skilled in the art understand the technical content of the present application, but it does not mean that the embodiments of the present application cannot be used in other devices, systems, environments or scenarios.

[0072] like Figure 1 As shown, the system architecture 100 according to this embodiment may include a first terminal device 101, a second terminal device 102, a third terminal device 103, a network 104, and a server 105. The network 104 is used to provide a medium for a communication link between the first terminal device 101, the second terminal device 102, the third terminal device 103, and the server 105. The network 104 may include various connection types, such as wired and / or wireless communication links, etc.

[0073] The user may use the first terminal device 101, the second terminal device 102, and the third terminal device 103 to interact with the server 105 through the network 104 to receive or send messages, etc. Various communication client applications may be installed on the first terminal device 101, the second terminal device 102, and the third terminal device 103, such as shopping applications, web browser applications, search applications, instant messaging tools, email clients, and / or social platform software, etc. (only for example).

[0074] The first terminal device 101, the second terminal device 102, and the third terminal device 103 may be various electronic devices having display screens and supporting web browsing, including but not limited to smart phones, tablet computers, laptop computers, desktop computers, and the like.

[0075] The server 105 may be a server that provides various services, such as a background management server (only as an example) that provides support for websites browsed by users using the first terminal device 101, the second terminal device 102, and the third terminal device 103. The background management server may analyze and process the received data such as user requests, and feed back the processing results (such as web pages, information, or data obtained or generated according to user requests) to the terminal device.

[0076] It should be noted that the server performance evaluation method provided in the embodiment of the present application can generally be performed by the server 105. Accordingly, the server performance evaluation device provided in the embodiment of the present application can generally be arranged in the server 105. The server performance evaluation method provided in the embodiment of the present application can also be performed by a server or server cluster that is different from the server 105 and can communicate with the first terminal device 101, the second terminal device 102, the third terminal device 103 and / or the server 105. Correspondingly, the server performance evaluation device provided in the embodiment of the present application can also be arranged in a server or server cluster that is different from the server 105 and can communicate with the first terminal device 101, the second terminal device 102, the third terminal device 103 and / or the server 105. Alternatively, the server performance evaluation method provided in the embodiment of the present application can also be performed by the first terminal device 101, the second terminal device 102 or the third terminal device 103, or can also be performed by other terminal devices that are different from the first terminal device 101, the second terminal device 102 or the third terminal device 103. Accordingly, the server performance evaluation device provided in the embodiment of the present application can also be set in the first terminal device 101, the second terminal device 102 or the third terminal device 103, or in other terminal devices different from the first terminal device 101, the second terminal device 102 or the third terminal device 103.

[0077] It should be understood that Figure 1 The number of terminal devices, networks and servers in the embodiment is only . According to the implementation requirements, there can be any number of terminal devices, networks and servers.

[0078] Figure 2 A flow chart of a server performance evaluation method according to an embodiment of the present application is shown.

[0079] like Figure 2 As shown, the server performance evaluation method includes operations S201 to S203.

[0080] In operation S201, in response to a transaction volume prediction instruction, target transaction information of a target product in an i-th time period is obtained, wherein the target transaction information is determined by a transaction knowledge graph constructed based on initial transaction information of the target product, and the transaction information represents a variety of initial transaction features of the target product in a transaction market.

[0081] In operation S202, the target transaction information is input into a target transaction volume prediction model, and the predicted transaction volume of the target product in the i+1th time period is output, wherein the predicted transaction volume represents the number of visits to the target product. The target transaction volume prediction model is obtained by training an initial recurrent neural network and an initial self-attention network using multiple target training transaction information.

[0082] In operation S203, a performance evaluation result is generated according to the predicted transaction volume and the device parameters of the target server storing the target product, wherein the performance evaluation result indicates whether the target server can bear the number of accesses in the (i+1)th time period.

[0083] According to an embodiment of the present application, the transaction volume prediction instruction can be input by a staff member on an electronic device such as a mobile phone or a computer, and the electronic device automatically generates the transaction volume prediction instruction in response to the operation, or the electronic device automatically generates the transaction volume prediction instruction at a preset time period (for example, one hour).

[0084] According to an embodiment of the present application, the i-th time period may be a historical time period, such as the previous hour, wherein the number of visits to the target product and the target transaction information within the i-th time period are known, and the i+1-th time period may be a historical time period or a future time period, such as the next hour.

[0085] According to the embodiment of the present application, the target product may be a product such as stocks, funds, precious metals, crude oil, etc., or operations such as checking balance, phone bills, services, watching advertisements, etc. The relevant product programs of the target product are stored in the target server for user access.

[0086] According to the embodiment of the present application, the transaction information may include various transaction features such as transaction code, transaction date, time, code transaction volume, transaction amount, product price change data, etc. Device parameters may refer to hardware parameters such as server capacity and bandwidth that affect user access quality.

[0087] According to an embodiment of the present application, before starting the server performance evaluation, a corresponding transaction knowledge graph is constructed through the initial transaction information of the target product containing multiple initial transaction features in the historical time period, and at least one target transaction feature is determined from the multiple initial transaction features based on the transaction knowledge graph. After the electronic device responds to the transaction volume prediction instruction, the electronic device automatically obtains the target transaction information of the target product containing at least one target transaction feature.

[0088] According to the embodiment of the present application, the target transaction information acquired in the i-th time period is input into the trained target transaction volume prediction model, so that the predicted transaction volume of the target product in the i+1-th time period can be obtained, and the predicted transaction volume represents the number of visits to the target product in the i+1-th time period. Based on the predicted transaction volume and the device parameters of the target server storing the target product, the performance evaluation result of the target product is obtained, and based on the performance evaluation result, it can be determined whether the target server storing the target product needs to be expanded or the bandwidth needs to be increased in advance. Maintenance.

[0089] According to the embodiment of the present application, by obtaining the target transaction information of the target product in the i-th time period, the target transaction information is input into the target transaction volume prediction model, the predicted transaction volume of the target product in the i+1-th time period is output, and the performance evaluation result is generated according to the predicted transaction volume and the device parameters of the target server storing the target product. Since the type of target transaction information is determined by the transaction knowledge graph constructed by the initial transaction information in the historical time period, the target transaction information with a greater impact relationship is input into the target transaction volume prediction model, which can reduce the data processing amount of the model, thereby reducing the computing resources used by the model. At the same time, the target transaction volume prediction model constructed based on the initial self-attention network can improve the prediction accuracy of the transaction volume, so that a more accurate server performance evaluation result can be obtained.

[0090] Figure 3 A schematic diagram of a transaction knowledge graph according to an embodiment of the present application is shown.

[0091] According to an embodiment of the present application, the target transaction information is obtained in the following manner: for any historical time period, a plurality of initial transaction features within the historical time period are obtained; a transaction knowledge graph of the target product is constructed based on the plurality of initial transaction features; and the target transaction information is determined based on the plurality of transaction knowledge graphs, wherein the target transaction information includes a plurality of target transaction features corresponding to the target product.

[0092] According to an embodiment of the present application, the historical time period may be a historical time period with an hourly gradient, such as the first hour before, the second hour before...the nth hour before, or a historical time period with a week-based gradient, such as the first week before, the second week before, and so on.

[0093] According to the embodiments of the present application, a knowledge graph is a technical method that uses a graph model to describe the relationship between knowledge and modeled objects. A knowledge graph is a graph-based data structure, in which nodes are called entities and edges are called relationships. The "entity-relationship-entity" composed of nodes and edges is the basic unit of the knowledge graph, i.e., a triple.

[0094] According to an embodiment of the present application, a plurality of initial transaction features in each historical time period corresponding to the target product are obtained, and a knowledge transaction graph of the historical time period is constructed based on the plurality of initial transaction features in the historical time period, such as Figure 3 An example is shown, and then the type of target transaction information related to the target product is determined based on multiple transaction knowledge graphs in multiple historical time periods.

[0095] According to an embodiment of the present application, by performing a transaction knowledge graph related to the historical time period for multiple initial transaction features of the target product, the type of target transaction information that affects the transaction volume of the target product is determined based on multiple transaction knowledge graphs of multiple historical time periods. This can reduce the amount of data processing for subsequent models, avoid the influence of irrelevant factors on the model prediction accuracy, and thus improve the model's calculation speed and prediction accuracy.

[0096] According to an embodiment of the present application, target transaction information is determined based on multiple transaction knowledge graphs, including: for each transaction knowledge graph, determining multiple intermediate transaction features related to the transaction volume of the target product from the transaction knowledge graph; for each intermediate transaction feature, calculating the ratio between the number of intermediate transaction features and the number of transaction knowledge graphs; when the ratio is greater than a preset threshold, determining the intermediate transaction feature as the target transaction information.

[0097] According to an embodiment of the present application, the preset threshold can be set specifically according to actual needs, for example, it can be set to 60%.

[0098] According to an embodiment of the present application, for each transaction knowledge graph, the intermediate transaction features that affect the transaction volume of the target product are determined from the transaction knowledge graph. For example, among 50 transaction knowledge graphs, the intermediate transaction features in 40 transaction knowledge graphs are transaction dates. In this case, the proportion of transaction dates is 80%.

[0099] According to an embodiment of the present application, since the proportion of transaction dates is 80%, which is greater than the preset threshold of 60%, the transaction date can be determined as a type of target transaction information.

[0100] According to an embodiment of the present application, by measuring the proportion of different intermediate transaction features in each transaction knowledge graph, when the proportion of each intermediate transaction feature is greater than a preset threshold, the intermediate transaction feature is determined as the target transaction feature, and the transaction features that have a smaller impact on the transaction volume of the target product can be eliminated, thereby reducing the data processing amount of the model and improving the prediction accuracy of the transaction volume of the target product.

[0101] According to an embodiment of the present application, before constructing a transaction knowledge graph, it also includes: for each initial transaction feature, preprocessing the initial transaction feature to obtain a preprocessed initial transaction feature, wherein the preprocessing includes removing at least one of missing values, outliers and duplicate values; normalizing multiple preprocessed initial transaction features to obtain normalized initial transaction features, so as to construct a transaction knowledge graph using the initial transaction features.

[0102] According to an embodiment of the present application, since the collected initial transaction features may contain some blank values, abnormal values, duplicate values, etc., and these problematic values ​​may affect the prediction accuracy of subsequent models, before constructing the transaction knowledge graph, these problematic values ​​are first cleaned up through preprocessing, such as deletion and other operations.

[0103] According to an embodiment of the present application, since different types of initial transaction features have different data scales after preprocessing, different types of preprocessed initial transaction features can be normalized to map data of different scales to a certain range (for example, [-1,1]). In this way, the dimensional differences between different features are eliminated, so that the model can learn and converge more efficiently during the training process.

[0104] Figure 4 A flowchart of a patrol method according to a target transaction volume prediction model in an embodiment of the present application is shown.

[0105] According to the embodiments of the present application, Figure 4 As shown, the training method of the target transaction volume prediction model includes operations S401 to S404.

[0106] In operation S401, in response to a model training instruction, a training sample set is obtained, wherein the training sample set includes a plurality of training transaction information corresponding to different training products in an i-th training time period and a transaction volume label corresponding to each training transaction information.

[0107] In operation S402, for each piece of training transaction information, the training transaction information is input into an initial transaction volume prediction model, and training prediction information is output, wherein the training prediction information represents the transaction volume prediction information of the training product in the i+1th training time period.

[0108] In operation S403, a target loss result is calculated according to the training prediction information and the transaction volume label corresponding to the training transaction information.

[0109] In operation S404, model parameters of the initial transaction volume prediction model are iteratively adjusted according to the target loss result to obtain a trained target transaction volume prediction model.

[0110] According to an embodiment of the present application, the type of the training product is the same as the type of the target product, the training transaction information is the same as the target transaction information, which is also determined by means of a transaction knowledge graph, and the transaction volume label may refer to the number of visits to the target product in the i+1th training time period. The training prediction information may refer to the predicted visit volume in the i+1th training time period.

[0111] According to an embodiment of the present application, the training transaction information of each i-th training time period is input into the initial transaction volume prediction model to predict the transaction volume prediction information of the training product in the i+1-th training time period, and the target loss result is calculated based on the transaction volume label and the training prediction information of the i+1-th training time period, wherein the target loss result can be obtained through the mean squared error (MSE) function, the cross entropy function, the logarithmic loss, etc. After obtaining the target loss result, the model parameters of the initial transaction volume prediction model are iteratively adjusted based on the target loss result, thereby obtaining a trained target transaction volume prediction model.

[0112] Figure 5 A data processing flow chart of an initial recurrent neural network according to an embodiment of the present application is shown.

[0113] According to an embodiment of the present application, the training transaction information includes a plurality of training transaction features.

[0114] According to an embodiment of the present application, training transaction information is input into an initial transaction volume prediction model, and training prediction information is output, including: for each training transaction information, multiple training transaction features are input into an initial recurrent neural network, and predicted transaction features corresponding to the training transaction information are output; and the predicted transaction features are subjected to self-attention weighted processing using an initial self-attention network to obtain training prediction information.

[0115] According to an embodiment of the present application, the initial recurrent neural network refers to a recurrent neural network (RNN), which is a type of recurrent neural network that takes sequence data as input, recurses in the direction of sequence evolution, and all nodes (recurrent units) are connected in a chain. Recurrent neural networks have memory, parameter sharing, and Turing completeness, so they have certain advantages when learning nonlinear features of sequences.

[0116] According to the embodiments of the present application, the Self-Attention Network Model is a model widely used in deep learning, especially in the Transformer architecture. The core idea of ​​the self-attention mechanism is to allow the model to pay attention to all other elements in the input sequence when processing each input element, so as to better capture the global dependencies in the input data.

[0117] According to an embodiment of the present application, for each training transaction information, multiple training transaction features of the training transaction information are input into the initial recurrent neural network, and the predicted transaction features corresponding to the training transaction information are output. t .

[0118] According to the embodiments of the present application, Figure 5 As shown, X t-1 is the input at time t-1 (i.e., the i-th time period), which can be either a vector or a matrix (i.e., a feature vector or feature matrix formed by multiple training transaction features of training transaction information). The initial recurrent neural network calculates the predicted transaction feature O based on formula (1). t :

[0119] S t =f(W x *X t +W s *S t-1 +B1)

[0120] O t =g(W0*S t +B2) (1)

[0121] Among them, W x is the weight matrix from the input layer to the hidden layer, W0 is the weight matrix from the hidden layer to the output layer, W s is the weight of the last value of the hidden layer as the input this time, S t is the hidden layer at time t, O t is the output at time t, f(), g() are activation functions

[0122] According to the embodiment of the present application, the collected multiple training transaction features (transaction code, transaction amount, amount involved, whether it is a holiday) are converted into feature vectors in the form of X as the input of the initial recurrent neural network. t , predict the predicted transaction features O for the next time period through the initial recurrent neural network t , and the hidden layer value S t Used to calculate the next hidden layer value.

[0123] According to the embodiment of the present application, the initial self-attention network is then used to process the predicted transaction feature O t , to output the training prediction information G through the learning mechanism inside the self-attention network t Among them, the self-attention network calculates the training prediction information G through formula (2) t :

[0124] G t =softmax(O t O t T / ) t (2)

[0125] Among them, softmax is the normalized exponential function, d is O t The matrix dimensions of the constructed matrix.

[0126] According to an embodiment of the present application, the training prediction information G t The actual trading volume label X for the next time period t+1 The target loss value can be obtained by calculation.

[0127] In a specific embodiment, when the transaction volume on the fund homepage surges, the transaction volume on the fund details page generally also surges. There is an implicit logical relationship between the two. The fund details page is entered only after entering the fund homepage. Adding a self-attention network can better fit the transaction volume fluctuations of this type of transactions.

[0128] According to an embodiment of the present application, the global dependencies in the input data are captured through a self-attention network, and self-attention weighted processing is performed based on predicted transaction features, so as to consider the training prediction information of the global dependencies, which can improve the prediction accuracy of the target transaction volume prediction model for the number of visits to the target product in the next time period.

[0129] According to an embodiment of the present application, the training transaction information is input into the initial transaction volume prediction model, and the training prediction information is output, which also includes:

[0130] When the initial recurrent neural network processes the training transaction information, the initial weight parameters of each initial neuron in the initial recurrent neural network are obtained; for each initial neuron, the initial weight parameters are processed by the gradient descent algorithm to obtain the initial gradient value; based on the initial gradient value, the initial weight parameters of the initial neuron are updated to obtain the updated initial recurrent neural network.

[0131] According to an embodiment of the present application, in the process of training an initial recurrent neural network, the initial weight parameters of each initial neuron in the initial recurrent neural network are obtained in real time, the initial weight parameters are processed using a gradient descent algorithm to calculate an initial gradient value, and the initial weight parameters of the initial neurons are updated based on the initial gradient value to obtain an updated initial recurrent neural network, so as to use the updated initial recurrent neural network to process the next training transaction information.

[0132] According to an embodiment of the present application, by using a gradient descent algorithm to update the initial weight parameters of the initial neurons, the target loss value during the training process can be gradually reduced, thereby helping the initial recurrent neural network to converge better, thereby improving the prediction accuracy of the target transaction volume prediction model.

[0133] According to an embodiment of the present application, the target transaction volume prediction model includes a target recurrent neural network and a target self-attention network.

[0134] According to an embodiment of the present application, the server performance evaluation method further includes:

[0135] The target weight parameters of each target neuron in the target recurrent neural network are obtained when the target transaction volume prediction model processes the target transaction information; for each target neuron, the target weight parameters are processed by the gradient descent algorithm to obtain the target gradient value; based on the target gradient value, the target weight parameters of the target neuron are updated to obtain an updated target recurrent neural network, wherein the updated target transaction volume prediction model includes the updated target recurrent neural network and the target self-attention network.

[0136] According to an embodiment of the present application, during the use of the target transaction volume prediction model, the target weight parameters of each target neuron in the target recurrent neural network can also be collected in real time when the target recurrent neural network processes the target transaction information, and the target weight parameters are processed again using the gradient descent algorithm to update the target weight parameters of the target neuron based on the obtained target gradient value, thereby obtaining an updated target recurrent neural network and an updated target transaction volume prediction model.

[0137] According to an embodiment of the present application, by reusing the gradient descent algorithm to update the initial weight parameters of the initial neurons during the use of the target transaction volume prediction model, the prediction accuracy of the target transaction volume prediction model can be further improved.

[0138] According to an embodiment of the present application, the server performance evaluation method further includes:

[0139] The actual number of visits in the i+1th time period is obtained; an optimized loss result is generated according to the actual number of visits and the predicted transaction volume; and an updated target transaction volume prediction model or a target transaction volume prediction model is optimized according to the optimized loss result to obtain a new target transaction volume prediction model.

[0140] According to an embodiment of the present application, after the time node reaches the i+1th time period, the actual number of visits to the target product in the i+1th time period is collected in a timely manner, and the optimized loss result is calculated based on the actual number of visits and the predicted transaction volume predicted by the target transaction volume prediction model, thereby optimizing the updated target transaction volume prediction model or the target transaction volume prediction model based on the optimized loss result, thereby obtaining a new target transaction volume prediction model.

[0141] According to an embodiment of the present application, by calculating the optimized loss result through the actual number of visits and the predicted transaction volume, and further optimizing the target transaction volume prediction model based on the optimized loss result, the prediction accuracy of the target transaction volume prediction model can be further improved.

[0142] Figure 6A block diagram of a server performance evaluation device according to an embodiment of the present application is shown.

[0143] like Figure 6 As shown, the server performance evaluation device 600 includes an acquisition module 610 , a prediction module 620 , and an evaluation module 630 .

[0144] The acquisition module 610 is used to respond to the transaction volume prediction instruction and obtain the target transaction information of the target product in the i-th time period, wherein the target transaction information is determined by the transaction knowledge graph constructed based on the initial transaction information of the target product, and the transaction information represents various initial transaction characteristics of the target product in the transaction market.

[0145] The prediction module 620 is used to input the target transaction information into the target transaction volume prediction model, and output the predicted transaction volume of the target product in the i+1th time period, wherein the predicted transaction volume represents the number of visits to the target product, and the target transaction volume prediction model is obtained by training the initial recurrent neural network and the initial self-attention network using multiple target training transaction information.

[0146] The evaluation module 630 is used to generate a performance evaluation result based on the predicted transaction volume and the device parameters of the target server storing the target product, wherein the performance evaluation result represents whether the target server can bear the number of visits in the i+1th time period.

[0147] According to the embodiment of the present application, by obtaining the target transaction information of the target product in the i-th time period, the target transaction information is input into the target transaction volume prediction model, the predicted transaction volume of the target product in the i+1-th time period is output, and the performance evaluation result is generated according to the predicted transaction volume and the device parameters of the target server storing the target product. Since the type of target transaction information is determined by the transaction knowledge graph constructed by the initial transaction information in the historical time period, the target transaction information with a greater impact relationship is input into the target transaction volume prediction model, which can reduce the data processing amount of the model, thereby reducing the computing resources used by the model. At the same time, the target transaction volume prediction model constructed based on the initial self-attention network can improve the prediction accuracy of the transaction volume, so that a more accurate server performance evaluation result can be obtained.

[0148] According to an embodiment of the present application, a device for generating target transaction information includes a second acquisition module, a construction module, and a determination module.

[0149] The second acquisition module is used to acquire multiple initial transaction features within any historical time period.

[0150] A construction module is used to construct a transaction knowledge graph of a target product based on multiple initial transaction features.

[0151] A determination module is used to determine target transaction information based on multiple transaction knowledge graphs, wherein the target transaction information includes multiple target transaction features corresponding to the target product.

[0152] According to an embodiment of the present application, the determination module includes a first determination unit, a calculation unit, and a second determination unit.

[0153] The first determination unit is used to determine, for each transaction knowledge graph, a plurality of intermediate transaction features related to the transaction volume of the target product from the transaction knowledge graph.

[0154] The calculation unit is used to calculate the ratio between the number of intermediate transaction features and the number of transaction knowledge graphs for each intermediate transaction feature.

[0155] The second determining unit is configured to determine the intermediate transaction feature as target transaction information when the proportion is greater than a preset threshold.

[0156] According to an embodiment of the present application, the device for generating target transaction information further includes a preprocessing module and a normalization module.

[0157] The preprocessing module is used to preprocess the initial transaction features for each initial transaction feature to obtain the preprocessed initial transaction features, wherein the preprocessing includes removing at least one of missing values, abnormal values ​​and duplicate values.

[0158] The normalization module is used to normalize multiple preprocessed initial transaction features to obtain normalized initial transaction features so as to construct a transaction knowledge graph using the initial transaction features.

[0159] According to an embodiment of the present application, a training device for a target transaction volume prediction model includes a third acquisition module, a second prediction module, a calculation module, and a training module.

[0160] The third acquisition module is used to obtain a training sample set in response to a model training instruction, wherein the training sample set includes a plurality of training transaction information corresponding to different training products in an i-th training time period and a transaction volume label corresponding to each training transaction information.

[0161] The second prediction module is used to input the training transaction information into the initial transaction volume prediction model for each training transaction information, and output the training prediction information, wherein the training prediction information represents the transaction volume prediction information of the training product in the i+1th training time period.

[0162] The calculation module is used to calculate the target loss result according to the training prediction information and the transaction volume label corresponding to the training transaction information.

[0163] The training module is used to iteratively adjust the model parameters of the initial transaction volume prediction model according to the target loss result to obtain a trained target transaction volume prediction model.

[0164] According to an embodiment of the present application, the training transaction information includes a plurality of training transaction features.

[0165] According to an embodiment of the present application, the second prediction module includes a prediction unit, a self-attention unit

[0166] The prediction unit is used to input multiple training transaction features into the initial recurrent neural network for each training transaction information, and output the predicted transaction features corresponding to the training transaction information.

[0167] The self-attention unit is used to perform self-attention weighted processing on the predicted transaction features using the initial self-attention network to obtain training prediction information.

[0168] According to an embodiment of the present application, the second prediction module further includes an acquisition unit, a gradient unit, and an update unit.

[0169] The acquisition unit is used to acquire the initial weight parameter of each initial neuron in the initial recurrent neural network when the initial recurrent neural network processes the training transaction information.

[0170] The gradient unit is used to process the initial weight parameters for each initial neuron using the gradient descent algorithm to obtain the initial gradient value.

[0171] The updating unit is used to update the initial weight parameters of the initial neurons based on the initial gradient values ​​to obtain an updated initial recurrent neural network.

[0172] According to an embodiment of the present application, the target transaction volume prediction model includes a target recurrent neural network and a target self-attention network.

[0173] According to an embodiment of the present application, the server performance evaluation device 600 also includes a fourth acquisition module, a gradient calculation module, and a weight updating module.

[0174] The fourth acquisition module is used to obtain the target weight parameter of each target neuron in the target recurrent neural network when the target transaction volume prediction model processes the target transaction information.

[0175] The gradient calculation module is used to process the target weight parameters for each target neuron using the gradient descent algorithm to obtain the target gradient value.

[0176] The weight update module is used to update the target weight parameters of the target neuron based on the target gradient value to obtain an updated target recurrent neural network, wherein the updated target transaction volume prediction model includes the updated target recurrent neural network and the target self-attention network.

[0177] According to an embodiment of the present application, the server performance evaluation device 600 further includes a fifth acquisition module, a generation module, and an optimization module.

[0178] A fifth acquisition module, used to acquire the actual number of visits in the i+1th time period;

[0179] A generation module, used to generate optimized loss results based on the actual number of visits and predicted transaction volume;

[0180] The optimization module is used to optimize the updated target transaction volume prediction model or the target transaction volume prediction model according to the optimization loss result to obtain a new target transaction volume prediction model.

[0181] According to the embodiments of the present application, any one or more of the modules, submodules, units, and subunits, or at least part of the functions of any one of them can be implemented in one module. According to the embodiments of the present application, any one or more of the modules, submodules, units, and subunits can be split into multiple modules for implementation. According to the embodiments of the present application, any one or more of the modules, submodules, units, and subunits can be at least partially implemented as hardware circuits, such as field programmable gate arrays (FPGAs), programmable logic arrays (PLAs), systems on chips, systems on substrates, systems on packages, application specific integrated circuits (ASICs), or can be implemented by hardware or firmware in any other reasonable way of integrating or packaging the circuit, or in any one of the three implementation methods of software, hardware, and firmware, or in any appropriate combination of any of them. Alternatively, according to the embodiments of the present application, one or more of the modules, submodules, units, and subunits can be at least partially implemented as computer program modules, and when the computer program modules are run, the corresponding functions can be executed.

[0182] For example, any multiple of the acquisition module 610, the prediction module 620, and the evaluation module 630 can be combined in one module / unit / sub-unit for implementation, or any one of the modules / units / sub-units can be split into multiple modules / units / sub-units. Alternatively, at least part of the functions of one or more of these modules / units / sub-units can be combined with at least part of the functions of other modules / units / sub-units and implemented in one module / unit / sub-unit. According to an embodiment of the present application, at least one of the acquisition module 610, the prediction module 620, and the evaluation module 630 can be at least partially implemented as a hardware circuit, such as a field programmable gate array (FPGA), a programmable logic array (PLA), a system on a chip, a system on a substrate, a system on a package, an application specific integrated circuit (ASIC), or can be implemented by hardware or firmware such as any other reasonable way of integrating or packaging the circuit, or implemented in any one of the three implementation methods of software, hardware, and firmware or in a suitable combination of any of them. Alternatively, at least one of the acquisition module 610 , the prediction module 620 , and the evaluation module 630 may be at least partially implemented as a computer program module, and when the computer program module is executed, the corresponding function may be executed.

[0183] It should be noted that the server performance evaluation device part in the embodiment of the present application corresponds to the server performance evaluation method part in the embodiment of the present application. The description of the server performance evaluation device part specifically refers to the server performance evaluation method part, which will not be repeated here.

[0184] Figure 7 A block diagram of an electronic device suitable for implementing the method described above according to an embodiment of the present application is shown. Figure 7 The electronic device shown is merely an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0185] like Figure 7 As shown, the electronic device 700 according to an embodiment of the present application includes a processor 701, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 702 or a program loaded from a storage part 708 to a random access memory (RAM) 703. The processor 701 may include, for example, a general-purpose microprocessor (such as a CPU), an instruction set processor and / or a related chipset and / or a special-purpose microprocessor (for example, an application-specific integrated circuit (ASIC)), etc. The processor 701 may also include an onboard memory for caching purposes. The processor 701 may include a single processing unit or multiple processing units for performing different actions of the method flow according to an embodiment of the present application.

[0186] In RAM 703, various programs and data required for the operation of electronic device 700 are stored. Processor 701, ROM 702 and RAM 703 are connected to each other via bus 704. Processor 701 performs various operations of the method flow according to the embodiment of the present application by executing the program in ROM 702 and / or RAM 703. It should be noted that the program can also be stored in one or more memories other than ROM 702 and RAM 703. Processor 701 can also perform various operations of the method flow according to the embodiment of the present application by executing the program stored in the one or more memories.

[0187] According to an embodiment of the present application, the electronic device 700 may further include an input / output (I / O) interface 705, which is also connected to the bus 704. The electronic device 700 may further include one or more of the following components connected to the input / output (I / O) interface 705: an input portion 706 including a keyboard, a mouse, etc.; an output portion 707 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage portion 708 including a hard disk, etc.; and a communication portion 709 including a network interface card such as a LAN card, a modem, etc. The communication portion 709 performs communication processing via a network such as the Internet. A drive 710 is also connected to the input / output (I / O) interface 705 as needed. A removable medium 711, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 710 as needed, so that a computer program read therefrom is installed into the storage portion 708 as needed.

[0188] According to an embodiment of the present application, the method flow according to the embodiment of the present application can be implemented as a computer software program. For example, an embodiment of the present application includes a computer program product, which includes a computer program carried on a computer-readable storage medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 709, and / or installed from the removable medium 711. When the computer program is executed by the processor 701, the above-mentioned functions defined in the system of the embodiment of the present application are executed. According to an embodiment of the present application, the system, equipment, device, module, unit, etc. described above can be implemented by a computer program module.

[0189] The present application also provides a computer-readable storage medium, which may be included in the device / apparatus / system described in the above embodiments; or may exist independently without being assembled into the device / apparatus / system. The above computer-readable storage medium carries one or more programs, and when the above one or more programs are executed, the method according to the embodiment of the present application is implemented.

[0190] According to an embodiment of the present application, the computer-readable storage medium may be a non-volatile computer-readable storage medium. For example, it may include, but is not limited to: a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In the present application, a computer-readable storage medium may be any tangible medium containing or storing a program that may be used by or in combination with an instruction execution system, apparatus, or device.

[0191] For example, according to an embodiment of the present application, the computer-readable storage medium may include the ROM 702 and / or the RAM 703 described above and / or one or more memories other than the ROM 702 and the RAM 703 .

[0192] An embodiment of the present application also includes a computer program product, which includes a computer program, which contains program code for executing the method provided by the embodiment of the present application. When the computer program product runs on an electronic device, the program code is used to enable the electronic device to implement the method provided by the embodiment of the present application.

[0193] When the computer program is executed by the processor 701, the above functions defined in the system / device of the embodiment of the present application are executed. According to the embodiment of the present application, the system, device, module, unit, etc. described above can be implemented by a computer program module.

[0194] In one embodiment, the computer program may rely on tangible storage media such as optical storage devices, magnetic storage devices, etc. In another embodiment, the computer program may also be transmitted and distributed in the form of signals on a network medium, and downloaded and installed through the communication part 709, and / or installed from the removable medium 711. The program code contained in the computer program may be transmitted using any appropriate network medium, including but not limited to: wireless, wired, etc., or any suitable combination of the above.

[0195] According to an embodiment of the present application, the program code for executing the computer program provided by the embodiment of the present application can be written in any combination of one or more programming languages, and specifically, these computing programs can be implemented using high-level process and / or object-oriented programming languages, and / or assembly / machine languages. Programming languages ​​include, but are not limited to, such as Java, C++, python, "C" language or similar programming languages. The program code can be executed entirely on the user computing device, partially on the user device, partially on the remote computing device, or completely on the remote computing device or server. In the case of a remote computing device, the remote computing device can be connected to the user computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computing device (e.g., using an Internet service provider to connect through the Internet).

[0196] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present application. In this regard, each box in the flowchart or block diagram may represent a module, a program segment, or a part of a code, and the above-mentioned module, program segment, or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box may also occur in a different order from the order marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they may sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flowchart, and the combination of the boxes in the block diagram or flowchart, can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions. It will be appreciated by those skilled in the art that the features recorded in the various embodiments of the present application can be combined and / or combined in a variety of ways, even if such a combination or combination is not clearly recorded in the present application. In particular, without departing from the spirit and teachings of the present application, the features described in the various embodiments of the present application may be combined and / or combined in a variety of ways. All of these combinations and / or combinations fall within the scope of the present application.

[0197] The embodiments of the present application are described above. However, these embodiments are only for the purpose of illustration, and are not intended to limit the scope of the present application. Although each embodiment is described above, this does not mean that the measures in each embodiment cannot be used in combination advantageously. The present application does not depart from the scope of the present application, and those skilled in the art may make various substitutions and modifications, which should all fall within the scope of the present application.

Claims

1. A server performance evaluation method, characterized in that: include: In response to the transaction volume prediction instruction, target transaction information of a target product in an i-th time period is obtained, wherein the target transaction information is determined by a transaction knowledge graph constructed based on initial transaction information of the target product, and the transaction information represents a plurality of initial transaction features of the target product in a transaction market; Input the target transaction information into a target transaction volume prediction model, and output the predicted transaction volume of the target product in the i+1th time period, wherein the predicted transaction volume represents the number of visits to the target product, and the target transaction volume prediction model is obtained by training an initial recurrent neural network and an initial self-attention network using multiple target training transaction information; A performance evaluation result is generated according to the predicted transaction volume and the device parameters of the target server storing the target product, wherein the performance evaluation result indicates whether the target server can bear the number of visits in the (i+1)th time period.

2. The method according to claim 1, characterized in that The target transaction information is obtained in the following manner: For any historical time period, obtaining a plurality of initial transaction features within the historical time period; Constructing the transaction knowledge graph of the target product according to a plurality of the initial transaction features; The target transaction information is determined according to the multiple transaction knowledge graphs, wherein the target transaction information includes multiple target transaction features corresponding to the target product.

3. The method according to claim 2, characterized in that Determining the target transaction information according to the plurality of transaction knowledge graphs includes: For each of the transaction knowledge graphs, determining a plurality of intermediate transaction features related to the transaction volume of the target product from the transaction knowledge graph; For each of the intermediate transaction features, calculating the ratio between the number of the intermediate transaction features and the number of the transaction knowledge graphs; When the proportion is greater than a preset threshold, the intermediate transaction feature is determined as the target transaction information.

4. The method according to claim 2, characterized in that: Before constructing the transaction knowledge graph, it also includes: For each of the initial transaction features, preprocessing the initial transaction features to obtain preprocessed initial transaction features, wherein the preprocessing includes removing at least one of missing values, abnormal values, and duplicate values; The plurality of preprocessed initial transaction features are normalized to obtain normalized initial transaction features, so as to construct the transaction knowledge graph using the initial transaction features.

5. The method according to claim 1, characterized in that The target transaction volume prediction model is trained in the following way: In response to the model training instruction, a training sample set is obtained, wherein the training sample set includes a plurality of training transaction information corresponding to different training products in the i-th training time period and a transaction volume label corresponding to each of the training transaction information; For each of the training transaction information, input the training transaction information into the initial transaction volume prediction model, and output training prediction information, wherein the training prediction information represents the transaction volume prediction information of the training product in the i+1th training time period; Calculating a target loss result according to the training prediction information and a transaction volume label corresponding to the training transaction information; The model parameters of the initial transaction volume prediction model are iteratively adjusted according to the target loss result to obtain a trained target transaction volume prediction model.

6. The method according to claim 5, characterized in that The training transaction information includes a plurality of training transaction features; The step of inputting the training transaction information into the initial transaction volume prediction model and outputting the training prediction information includes: For each of the training transaction information, input a plurality of the training transaction features into an initial recurrent neural network, and output a predicted transaction feature corresponding to the training transaction information; The predicted transaction features are subjected to self-attention weighted processing using an initial self-attention network to obtain the training prediction information.

7. The method according to claim 6, characterized in that Also includes: In the case where the initial recurrent neural network processes the training transaction information, obtaining an initial weight parameter of each initial neuron in the initial recurrent neural network; For each of the initial neurons, the initial weight parameters are processed using a gradient descent algorithm to obtain an initial gradient value; The initial weight parameters of the initial neurons are updated based on the initial gradient values ​​to obtain an updated initial recurrent neural network.

8. The method according to any one of claims 1 to 4, characterized in that The target transaction volume prediction model includes a target recurrent neural network and a target self-attention network; Wherein, the server performance evaluation method further includes: Obtaining a target weight parameter of each target neuron in the target recurrent neural network when the target transaction volume prediction model processes the target transaction information; For each of the target neurons, the target weight parameter is processed using a gradient descent algorithm to obtain a target gradient value; The target weight parameters of the target neuron are updated based on the target gradient value to obtain an updated target recurrent neural network, wherein the updated target transaction volume prediction model includes the updated target recurrent neural network and the target self-attention network.

9. The method according to claim 8, characterized in that Also includes: Obtaining the actual number of visits in the i+1th time period; Generate an optimized loss result according to the actual number of visits and the predicted transaction volume; The updated target transaction volume prediction model or the target transaction volume prediction model is optimized according to the optimization loss result to obtain a new target transaction volume prediction model.

10. A server performance evaluation device, comprising: an acquisition module, configured to acquire target transaction information of a target product in an i-th time period in response to a transaction volume prediction instruction, wherein the target transaction information is determined by a transaction knowledge graph constructed based on initial transaction information of the target product, and the transaction information represents a plurality of initial transaction features of the target product in a transaction market; A prediction module, used to input the target transaction information into a target transaction volume prediction model, and output the predicted transaction volume of the target product in the i+1th time period, wherein the predicted transaction volume represents the number of visits to the target product, and the target transaction volume prediction model is obtained by training an initial recurrent neural network and an initial self-attention network using multiple target training transaction information; An evaluation module is used to generate a performance evaluation result based on the predicted transaction volume and the device parameters of the target server storing the target product, wherein the performance evaluation result represents whether the target server can bear the number of visits in the i+1th time period.

11. An electronic device, comprising: one or more processors; a memory for storing one or more programs, When the one or more programs are executed by the one or more processors, the one or more processors implement the method according to any one of claims 1 to 9.

12. A computer-readable storage medium having executable instructions stored thereon, which, when executed by a processor, enables the processor to implement the method according to any one of claims 1 to 9.

13. A computer program product comprising a computer program, which, when executed by a processor, implements the method according to any one of claims 1 to 9.

Images