A data processing method, device, equipment, storage medium and program product

Abstract

The present application provides a data processing method, device, equipment, storage medium and program product, which are applied to various scenarios such as cloud technology, artificial intelligence and games; the data processing method includes: displaying a virtual scene, responding to a data processing operation for the virtual scene, obtaining virtual activity data, wherein the virtual activity data is data associated with an interactive operation in the virtual scene; based on the virtual activity data, determining at least two virtual events corresponding to each specified event type, wherein the specified event type is a virtual event type in the virtual scene; determining a prediction result based on at least two virtual events, and determining the event contribution of each virtual event to the prediction result based on at least two virtual events, and obtaining at least two event contributions corresponding to at least two virtual events; displaying the prediction result and at least two event contributions corresponding to at least two virtual events. Through the present application, the effect of data processing can be improved.

Description

Technical Field

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

[0002] With the rapid development of artificial intelligence, artificial intelligence has been applied in various aspects; for example, in virtual scenes, artificial intelligence is often used to process data in virtual scenes, and the obtained results are displayed to prompt users or perform targeted processing.

[0003] However, in the above processing of data in the virtual scene, the obtained results usually include prediction results and events related to the prediction results. Therefore, the granularity of the obtained results is coarse and the data processing effect is poor. Summary of the invention

[0004] The embodiments of the present application provide a data processing method, apparatus, device, computer-readable storage medium, and computer program product, which can indicate the effect of data processing.

[0005] The technical solution of the embodiment of the present application is implemented as follows:

[0006] The present application provides a data processing method, including:

[0007] Displaying a virtual scene, and acquiring virtual activity data in response to a data processing operation on the virtual scene, wherein the virtual activity data is data associated with an interactive operation in the virtual scene;

[0008] Based on the virtual activity data, determining at least two virtual events corresponding to respective designated event types, wherein the designated event types are virtual event types in the virtual scene;

[0009] Determine a prediction result based on at least two of the virtual events, and determine an event contribution of each of the virtual events to the prediction result based on the at least two virtual events, to obtain at least two event contributions corresponding to the at least two virtual events;

[0010] The prediction result and at least two event contributions corresponding to at least two of the virtual events are displayed.

[0011] The present application provides a data processing device, including:

[0012] A data acquisition module, used to display a virtual scene, and acquire virtual activity data in response to a data processing operation on the virtual scene, wherein the virtual activity data is data associated with an interactive operation in the virtual scene;

[0013] An event determination module, configured to determine at least two virtual events corresponding to respective specified event types based on the virtual activity data, wherein the specified event types are virtual event types in the virtual scene;

[0014] A data processing module, used to determine a prediction result based on at least two of the virtual events, and determine an event contribution of each of the virtual events to the prediction result based on the at least two virtual events, to obtain at least two event contributions corresponding to the at least two virtual events;

[0015] The result display module is used to display the prediction result and at least two event contributions corresponding to at least two virtual events.

[0016] In an embodiment of the present application, the data processing module is also used to determine multiple virtual event sequences including a target virtual event based on at least two of the virtual events, wherein the target virtual event is any one of the at least two virtual events; obtain the sequence contribution of the target virtual event in each of the virtual event sequences, and obtain multiple sequence contributions corresponding to the multiple virtual event sequences; and determine the average of the multiple sequence contributions as the event contribution of the target virtual event to the prediction result.

[0017] In an embodiment of the present application, the data processing module is also used to perturb at least two of the virtual events based on a target virtual event among the at least two virtual events to obtain multiple sets of perturbed virtual events; train a fitting model based on the multiple sets of perturbed virtual events; and determine the event contribution of the target virtual event to the prediction result based on the fitting model.

[0018] In an embodiment of the present application, the data processing module is also used to extract features from each of the virtual events to obtain virtual event features; determine sub-prediction results based on the virtual event features; obtain the target time weight corresponding to each of the virtual events from the specified time weight corresponding to each of the specified event types, wherein the specified time weight is the correspondence between the event occurrence time and the importance; determine the fusion result of the target time weight and the sub-prediction results as the prediction result to be integrated, and obtain at least two prediction results to be integrated corresponding to at least two of the virtual events; integrate at least two of the prediction results to be integrated to obtain the prediction result.

[0019] In an embodiment of the present application, the data processing module is also used to determine the event change information corresponding to each of the virtual events based on a specified feature pool, wherein the specified feature pool includes historical virtual events corresponding to each historical time; and perform feature extraction on the virtual events and the event change information to obtain the virtual event features.

[0020] In the embodiment of the present application, the data processing module is further used to store each of the virtual events and the time corresponding to the virtual events in the designated feature pool.

[0021] In an embodiment of the present application, the data processing module is also used to determine the virtual event that belongs to numerical data from at least two of the virtual events to obtain a virtual event to be processed; normalize the virtual event to be processed to obtain a virtual event to be replaced; and replace the corresponding virtual event in at least two of the virtual events based on the virtual event to be replaced to obtain at least two first virtual events to be predicted.

[0022] In an embodiment of the present application, the data processing module is further used to determine the prediction result based on at least two of the first virtual events to be predicted.

[0023] In an embodiment of the present application, the data processing module is also used to determine the virtual event belonging to category data from at least two of the virtual events to obtain the virtual event to be converted; obtain the vector information corresponding to the virtual event to be converted to obtain a virtual event vector; based on the virtual event vector, replace the corresponding virtual event in at least two of the virtual events to obtain at least two second virtual events to be predicted.

[0024] In an embodiment of the present application, the data processing module is further used to determine the prediction result based on at least two of the second virtual events to be predicted.

[0025] In an embodiment of the present application, the prediction result is obtained through a result prediction model, and the data processing device also includes a model training module, which is used to determine the result prediction model to be trained based on the prediction result type, wherein the prediction result type includes at least one of a regression type and a classification type; based on the result prediction model to be trained, the feature training sample is predicted to obtain an estimated prediction result; based on the loss between the estimated prediction result and the prediction result label, the result prediction model to be trained is trained to obtain the result prediction model, wherein the prediction result label is the label of the feature training sample.

[0026] In an embodiment of the present application, the event contribution is obtained through a contribution prediction model, and the model training module is also used to predict the contribution training sample based on the contribution prediction model to be trained to obtain an estimated contribution; based on the loss between the estimated contribution and the contribution label, the contribution prediction model to be trained is trained to obtain the contribution prediction model, wherein the contribution label is the label of the contribution training sample.

[0027] In an embodiment of the present application, the data processing module is also used to obtain new virtual activity data when an update time arrives, wherein the update time is determined based on an update cycle; determine a new prediction result based on at least two new virtual events determined by the new virtual activity data, and determine a new event contribution of each of the new virtual events to the new prediction result, and obtain at least two new event contributions corresponding to at least two new virtual events; display the new prediction result and at least two new event contributions corresponding to at least two new virtual events.

[0028] In an embodiment of the present application, the data processing module is also used to perform a target interaction operation in the virtual scene in response to the interactive operations received for the displayed prediction results and at least two event contributions corresponding to at least two of the virtual events, wherein the target interaction operation is an interactive operation associated with the virtual event with a contribution less than a specified degree.

[0029] The present application provides a data processing device, including:

[0030] A memory for storing executable instructions;

[0031] The processor is used to implement the data processing method provided in the embodiment of the present application when executing the executable instructions stored in the memory.

[0032] An embodiment of the present application provides a computer-readable storage medium storing executable instructions, wherein the executable instructions are used to implement the data processing method provided in the embodiment of the present application when executed by a processor.

[0033] An embodiment of the present application provides a computer program product, including a computer program or instructions. When the computer program or instructions are executed by a processor, the data processing method provided by the embodiment of the present application is implemented.

[0034] The embodiments of the present application have at least the following beneficial effects: at least two virtual events are determined by collecting virtual activity data in a virtual scene, and a prediction result is determined based on the at least two virtual events, and the event contribution (i.e., the degree of influence) of each relevant virtual event to the prediction result is attributed; so that not only at least two virtual events associated with the prediction result can be obtained, but also the degree of influence of each virtual event on the prediction result can be obtained; thereby, the granularity of the processing results can be refined and the effect of data processing can be improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 is an exemplary schematic diagram of data processing results;

[0036] Figure 2 It is a schematic diagram of the architecture of the data processing system provided in the embodiment of the present application;

[0037] Figure 3 The embodiment of this application provides Figure 2 A schematic diagram of the composition structure of the terminal in FIG.

[0038] Figure 4 It is a flowchart of a data processing method provided in an embodiment of the present application;

[0039] Figure 5 is a schematic diagram of an exemplary data processing method provided in an embodiment of the present application;

[0040] Figure 6 This is a schematic diagram of the process of obtaining event contribution provided in the embodiment of the present application. Figure 1 ;

[0041] Figure 7 This is a schematic diagram of the process of obtaining event contribution provided in the embodiment of the present application. Figure 2 ;

[0042] Figure 8 is a schematic diagram of an exemplary data processing method provided in an embodiment of the present application;

[0043] Fig. 9 It is an exemplary data processing diagram provided in an embodiment of the present application. DETAILED DESCRIPTION

[0044] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings. The described embodiments should not be regarded as limiting the present application. All other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of this application.

[0045] In the following description, reference is made to “some embodiments”, which describe a subset of all possible embodiments, but it will be understood that “some embodiments” may be the same subset or different subsets of all possible embodiments and may be combined with each other without conflict.

[0046] In the following description, the terms "first\second" involved are merely used to distinguish similar objects and do not represent a specific ordering of the objects. It can be understood that "first\second" can be interchanged with a specific order or sequence where permitted, so that the embodiments of the present application described herein can be implemented in an order other than that illustrated or described herein.

[0047] Unless otherwise defined, all technical and scientific terms used in the embodiments of the present application have the same meaning as those commonly understood by those skilled in the art of the present application. The terms used in the embodiments of the present application are only for the purpose of describing the embodiments of the present application and are not intended to limit the present application.

[0048] Before further describing the embodiments of the present application in detail, the nouns and terms involved in the embodiments of the present application are explained. The nouns and terms involved in the embodiments of the present application are subject to the following interpretations.

[0049] 1) Artificial Intelligence (AI) is the theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results.

[0050] 2) Machine Learning (ML) is a multi-disciplinary subject involving probability theory, statistics, approximation theory, convex analysis, algorithm complexity theory and other disciplines. It specializes in studying how computers simulate or implement human learning behavior to acquire new knowledge or skills; reorganize existing knowledge structures to continuously improve their performance. Machine learning is the core of artificial intelligence and the fundamental way to make computers intelligent. Its applications are spread across all areas of artificial intelligence. Machine learning usually includes technologies such as artificial neural networks, belief networks, reinforcement learning, transfer learning and inductive learning.

[0051] 3) Artificial neural network is a mathematical model that imitates the structure and function of biological neural networks. The exemplary structures of artificial neural networks in the embodiments of the present application include deep neural networks (DNN), convolutional neural networks (CNN), recurrent neural networks (RNN), neural state machines (NSM), and phase-function neural networks (PFNN). The result prediction model and contribution prediction model involved in the embodiments of the present application are the models corresponding to artificial neural networks.

[0052] It should be noted that, in order to process the data in the virtual scene, the data in the virtual scene can be processed by artificial intelligence, however, the obtained results usually include prediction results and events related to the prediction results. Figure 1 , Figure 1 is an exemplary schematic diagram of data processing results; Figure 1 As shown, page 1-1 shows a virtual scene 1-11, and also shows a data processing result 1-12 for the virtual scene 1-11; however, the data processing result 1-12 includes the predicted winning rate of the two teams (the blue team has a winning rate of 98.5%, and the red team has a winning rate of 1.5%) and game events (the blue team leads in virtual economy, the blue team leads in virtual scoring, and the red team has a virtual lineup advantage). Therefore, the result obtained is a high-level global analysis, lacks quantitative data, and has a coarse granularity, so the data processing effect is poor.

[0053] In addition, in order to process the data in the virtual scene, a weight can be set in advance for each virtual event, and the game events occurring in the virtual scene can be sorted according to the weight to determine the importance of each game event; however, since the importance of each game event is static, and the importance of the same game event is different at different times, the importance of each game event is relative. Determining the importance of each game event based on the pre-set weight has low accuracy, is inconsistent with the actual situation, and has poor practicality.

[0054] Based on this, the embodiments of the present application provide a data processing method, apparatus, device, computer-readable storage medium and computer program product, which can improve the data processing effect. The following describes an exemplary application of the data processing device provided by the embodiments of the present application. The data processing device provided by the embodiments of the present application can be implemented as various types of terminals such as smart phones, smart watches, laptops, tablet computers, desktop computers, smart home appliances, set-top boxes, smart car devices, portable music players, personal digital assistants, dedicated messaging devices, intelligent voice interaction devices, portable gaming devices and smart speakers, and can also be implemented as servers. Below, an exemplary application of the data processing device when it is implemented as a terminal will be described.

[0055] See also Figure 2 , Figure 2 Schematic diagram of the data processing system provided in the embodiment of the present application; Figure 2 As shown, to support a data processing application, in the data processing system 100, the terminal 200 (the terminal 200-1 and the terminal 200-2 are shown as examples, referred to as data processing devices) is connected to the server 400 via the network 300. The network 300 can be a wide area network or a local area network, or a combination of the two. In addition, the data processing system 100 also includes a database 500 for providing data support to the server 400; and, Figure 2 What is shown in FIG. 5 is a case where the database 500 is independent of the server 400. In addition, the database 500 may also be integrated in the server 400, which is not limited in the embodiment of the present application.

[0056] Terminal 200 is used to display a virtual scene, and in response to a data processing operation for the virtual scene, sends a data processing request to server 400 through network 300; receives prediction results and at least two event contributions corresponding to at least two virtual events sent by server 400 through network 300, and displays the prediction results and at least two event contributions corresponding to at least two virtual events (graphic interface 210-1 and graphical interface 210-2 are shown as examples).

[0057] The server 400 is used to receive a data processing request sent by the terminal 200 through the network 300, and obtain virtual activity data in response to the data processing request, wherein the virtual activity data is data associated with interactive operations in the virtual scene; based on the virtual activity data, determine at least two virtual events corresponding to each specified event type, wherein the specified event type is a virtual event type in the virtual scene; determine a prediction result based on the at least two virtual events, and determine an event contribution of each virtual event to the prediction result based on the at least two virtual events, to obtain at least two event contributions corresponding to the at least two virtual events; and send the prediction result and the at least two event contributions corresponding to the at least two virtual events to the terminal 200 through the network 300.

[0058] In some embodiments, the server 400 may be an independent physical server, or a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content distribution networks (CDN, Content Delivery Network), and big data and artificial intelligence platforms. Terminal 200 may be a smart phone, smart watch, laptop, tablet computer, desktop computer, smart TV, set-top box, smart car device, portable music player, personal digital assistant, dedicated messaging device, portable game device and smart speaker, etc., but is not limited thereto. The terminal and the server may be directly or indirectly connected by wired or wireless communication, which is not limited in the embodiments of the present application.

[0059] See also Figure 3 , Figure 3 The embodiment of this application provides Figure 2 The schematic diagram of the terminal structure in FIG. Figure 3 The terminal 200 shown includes: at least one processor 210, a memory 250, at least one network interface 220 and a user interface 230. The various components in the terminal 200 are coupled together via a bus system 240. It is understood that the bus system 240 is used to achieve connection and communication between these components. In addition to the data bus, the bus system 240 also includes a power bus, a control bus and a status signal bus. However, for the sake of clarity, the bus system 240 is not described in detail. Figure 3 Various buses are labeled as bus system 240 .

[0060] The processor 210 can be an integrated circuit chip with signal processing capabilities, such as a general-purpose processor, a digital signal processor (DSP), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc., where the general-purpose processor can be a microprocessor or any conventional processor, etc.

[0061] The user interface 230 includes one or more output devices 231 that enable presentation of media content, including one or more speakers and / or one or more visual display screens. The user interface 230 also includes one or more input devices 232, including user interface components that facilitate user input, such as a keyboard, mouse, microphone, touch screen display, camera, other input buttons and controls.

[0062] The memory 250 may be removable, non-removable, or a combination thereof. Exemplary hardware devices include solid-state memory, hard disk drives, optical disk drives, etc. The memory 250 may optionally include one or more storage devices that are physically remote from the processor 210.

[0063] The memory 250 includes a volatile memory or a non-volatile memory, and may also include both volatile and non-volatile memories. The non-volatile memory may be a read-only memory (ROM), and the volatile memory may be a random access memory (RAM). The memory 250 described in the embodiments of the present application is intended to include any suitable type of memory.

[0064] In some embodiments, memory 250 can store data to support various operations, examples of which include programs, modules, and data structures, or a subset or superset thereof, as exemplarily described below.

[0065] Operating system 251, including system programs for processing various basic system services and performing hardware-related tasks, such as a framework layer, a core library layer, a driver layer, etc., for implementing various basic services and processing hardware-based tasks;

[0066] A network communication module 252, used to reach other computer devices via one or more (wired or wireless) network interfaces 220, exemplary network interfaces 220 include: Bluetooth, Wireless Fidelity (Wi-Fi), and Universal Serial Bus (USB);

[0067] a presentation module 253 for enabling presentation of information via one or more output devices 231 (e.g., display screen, speaker, etc.) associated with the user interface 230 (e.g., a user interface for operating peripherals and displaying content and information);

[0068] The input processing module 254 is used to detect one or more user inputs or interactions from one of the one or more input devices 232 and translate the detected inputs or interactions.

[0069] In some embodiments, the data processing device provided in the embodiments of the present application can be implemented in software. Figure 3 The data processing device 255 stored in the memory 250 is shown, which can be software in the form of a program and a plug-in, etc., including the following software modules: a data acquisition module 2551, an event determination module 2552, a data processing module 2553, a result display module 2554 and a model training module 2555. These modules are logical, so they can be arbitrarily combined or further split according to the functions implemented. The functions of each module will be explained below.

[0070] In some embodiments, the data processing device provided in the embodiments of the present application can be implemented in hardware. As an example, the data processing device provided in the embodiments of the present application can be a processor in the form of a hardware decoding processor, which is programmed to execute the data processing method provided in the embodiments of the present application. For example, the processor in the form of a hardware decoding processor can adopt one or more application-specific integrated circuits (ASICs), DSPs, programmable logic devices (PLDs), complex programmable logic devices (CPLDs), field programmable gate arrays (FPGAs), or other electronic components.

[0071] In some embodiments, the terminal or server can implement the data processing method provided in the embodiments of the present application by running a computer program. For example, the computer program can be a native program or software module in the operating system; it can be a native application (APP, Application), that is, a program that needs to be installed in the operating system to run, such as a live broadcast APP or an instant messaging APP; it can also be a small program, that is, a program that can be run only by downloading it to a browser environment; it can also be a small program that can be embedded in any APP. In short, the above-mentioned computer program can be any form of application, module or plug-in.

[0072] Below, the data processing method provided by the embodiment of the present application will be described in combination with the exemplary application and implementation of the data processing device provided by the embodiment of the present application. In addition, the data processing method provided by the embodiment of the present application is applied to various scenarios such as cloud technology, artificial intelligence, smart transportation and vehicle-mounted.

[0073] See also Figure 4 , Figure 4 is a flow chart of the data processing method provided in the embodiment of the present application, which will be combined with Figure 4 The steps shown are explained.

[0074] Step 401: Display a virtual scene, and acquire virtual activity data in response to a data processing operation on the virtual scene.

[0075] In the embodiment of the present application, when the data processing device runs an application for rendering a virtual scene, the virtual scene is displayed, wherein the displayed virtual scene includes data associated with the interactive operation. Then, when a data processing operation is triggered for the displayed virtual scene, the data processing device receives the data processing operation for the virtual scene. At this time, in response to the data processing operation, the data associated with the interactive operation included in the current virtual scene is obtained, and the virtual activity data is obtained.

[0076] It should be noted that the virtual activity data is the data associated with the interactive operations in the virtual scene, is the latest activity data in the virtual scene, and is real-time data; for example, the current virtual score, the current virtual kill volume, the current virtual economy, the current virtual lineup, the current defense resources, the current virtual supplies, etc. In addition, the data processing operation can be obtained by receiving input instructions, such as keyboard input, voice input, gesture input, etc.; it can also be obtained by detecting the trigger operation of the specified area, such as clicking the "data" button, etc.; the embodiments of the present application are not limited to this. In addition, the virtual activity data may be virtual game data, or it may not be virtual game data, wherein the virtual game data refers to the relevant data of the virtual object queue game, such as game game data. Here, the virtual scene is, for example, a game scene, and at this time, the virtual activity data is real-time game data (for example, real-time match data).

[0077] Step 402: Determine at least two virtual events corresponding to each specified event type based on the virtual activity data.

[0078] In the embodiment of the present application, each designated event type is set in the data processing device, or the data processing device can obtain each designated event type from other devices (e.g., a database device, a device that interacts with the data processing device), and the designated event type is a virtual event type in a virtual scene, corresponding to a virtual domain; thus, the data processing device divides the virtual activity data into types based on each designated event type, and obtains at least two virtual events. Here, the virtual event is, for example, a game event.

[0079] It should be noted that each of the at least two virtual events belongs to a specified event type, and the two virtual events can be of the same specified event type or of different specified event types. In addition, each specified event type is a pre-set virtual event type associated with the content to be predicted, including at least one of virtual transfer resources, virtual attack output, virtual defense resources, virtual regional resources, virtual game resources, virtual lineup resources, virtual field of view, virtual sound type, virtual position type, virtual equipment resources, and virtual scene type.

[0080] Step 403: determine a prediction result based on at least two virtual events, and determine an event contribution of each virtual event to the prediction result based on the at least two virtual events, to obtain at least two event contributions corresponding to the at least two virtual events.

[0081] In an embodiment of the present application, a data processing device performs a target prediction task based on features corresponding to at least two virtual events, thereby determining a prediction result; and after obtaining the prediction result, the data processing device also attributes the prediction result to determine the event contribution of each virtual event to the prediction result, thereby obtaining at least two event contributions for at least two virtual events.

[0082] It should be noted that at least two virtual events correspond to at least two event contributions one by one; each event contribution indicates the degree of influence of the virtual event corresponding to the event contribution on the prediction result; and the event contribution can be a scalar or a vector; when the event contribution is a scalar, if the event contribution is greater than the specified contribution (for example, 0), it indicates that the virtual event has a positive effect on the prediction result, and if the event contribution is less than the specified contribution, it indicates that the virtual event has a negative effect on the prediction result; when the event contribution is a vector, if the direction of the event contribution is consistent with the direction of the prediction result, it indicates that the virtual event has a positive effect on the prediction result, and the direction of the event contribution is opposite to the direction of the prediction result, it indicates that the virtual event has a negative effect on the prediction result. In addition, when the virtual acquired data is virtual game data, the prediction result can be the game result of each virtual game party, such as a virtual queue of victory or defeat, a virtual score, etc.

[0083] Here, the data processing device can implement the prediction results and the determination of at least two event contributions corresponding to at least two virtual events in itself; it can also send a data processing request carrying at least two virtual events to the server device, so that the server device determines the prediction results and the at least two event contributions corresponding to at least two virtual events based on the at least two virtual events, and receives the prediction results and the at least two event contributions corresponding to at least two virtual events sent by the server device; and so on. The embodiments of the present application are not limited to this.

[0084] Step 404: display the prediction result and at least two event contributions corresponding to at least two virtual events.

[0085] In an embodiment of the present application, after obtaining the prediction result and at least two event contributions corresponding to at least two virtual events, the data processing device will also display the prediction result and at least two event contributions corresponding to at least two virtual events.

[0086] It should be noted that the display of the prediction results and the contribution of at least two events corresponding to at least two virtual events can be simultaneous with the display of the virtual scene, and can also be switched with the display of the virtual scene, etc., and the embodiments of the present application are not limited to this.

[0087] For example, see Figure 5 , Figure 5 is a schematic diagram of an exemplary data processing method provided in an embodiment of the present application; Figure 5 As shown, page 5-1 displays a virtual scene 5-11 and a tab 5-12; when a click operation (called a data processing operation) is received for tab 5-12, the prediction result (the real-time winning rate of the blue team is 68%) and the attribution result (the event contributions corresponding to virtual events 5-13 to 5-18 are 39%, 18%, 12%, 3%, -1% and -3%, respectively, which are called at least two event contributions corresponding to at least one virtual event) are displayed.

[0088] It can be understood that by collecting virtual activity data in a virtual scene to determine at least two virtual events, and determining a prediction result based on the at least two virtual events, and attributing the event contribution of each relevant virtual event to the prediction result; not only can at least two virtual events associated with the prediction result be obtained, but also the degree of influence of each virtual event on the prediction result can be obtained; thereby, the granularity of the processing results can be refined, the comprehensiveness of the data display can be improved, and the effect of data processing can be improved.

[0089] See also Figure 6 , Figure 6This is a schematic diagram of the process of obtaining event contribution provided in the embodiment of the present application. Figure 1 ;like Figure 6 As shown, in an embodiment of the present application, in step 403, the data processing device determines the event contribution of each virtual event to the prediction result based on at least two virtual events, including steps 4031 to 4033, and each step is described below.

[0090] Step 4031: Based on at least two virtual events, determine multiple virtual event sequences including a target virtual event.

[0091] In an embodiment of the present application, a data processing device traverses at least two virtual events, determines the traversed virtual event as a target virtual event, and obtains, for the target virtual event, a plurality of virtual event sequences in which the target virtual event appears in different orders in the at least two virtual events, wherein the target virtual event is any one of the at least two virtual events.

[0092] It should be noted that the data processing device can also filter out multiple virtual events from at least two virtual events, and then obtain multiple virtual event sequences in which the target virtual event appears in different orders in the filtered multiple virtual events; wherein the screening method can be random, or can be based on a preset importance, etc., which is not limited to the embodiments of the present application.

[0093] Exemplarily, the multiple virtual event sequences may be a full arrangement of at least two virtual events, or may be a full arrangement of multiple selected virtual events.

[0094] Step 4032: Obtain the sequence contribution of the target virtual event in each virtual event sequence, and obtain multiple sequence contributions corresponding to the multiple virtual event sequences.

[0095] In the embodiment of the present application, the data processing device obtains the influence of the target virtual event on the prediction result when it appears in the corresponding order of appearance in each virtual event sequence, and thus obtains the sequence contribution, thereby obtaining multiple sequence contribution corresponding to multiple virtual event sequences. Among them, multiple virtual event sequences correspond to multiple sequence contribution one by one.

[0096] Step 4033: Determine the average of the contribution of the multiple sequences as the event contribution of the target virtual event to the prediction result.

[0097] In an embodiment of the present application, a data processing device calculates the average of the contribution of multiple sequences, and determines the calculated average of the contribution of multiple sequences as the event contribution of the target virtual event to the prediction result, thereby obtaining the event contribution of each pseudo event to the prediction result.

[0098] It can be understood that by integrating multiple virtual event sequences in which the target virtual event appears in different orders, as well as the sequence contribution of the target virtual event in each virtual event sequence, it is possible to accurately attribute the prediction results and at least one virtual event, thereby improving the granularity of the data processing results of the virtual scene.

[0099] See also Figure 7 , Figure 7 This is a schematic diagram of the process of obtaining event contribution provided in the embodiment of the present application. Figure 2 ;like Figure 7 As shown, in an embodiment of the present application, in step 403, the data processing device determines the event contribution of each virtual event to the prediction result based on at least two virtual events, including steps 4034 to 4036, and each step is described below.

[0100] Step 4034: Based on the target virtual event among the at least two virtual events, perturb the at least two virtual events to obtain a plurality of perturbed virtual event sets.

[0101] In the embodiment of the present application, the data processing device can select multiple virtual events including the target virtual event from at least two virtual events, and combine them into a disturbance virtual event set. After multiple selections, multiple disturbance virtual event sets are obtained. In addition, the disturbance can also be a process of changing the information corresponding to the virtual event; etc., which is not limited in the embodiment of the present application.

[0102] Step 4035: Train a fitting model based on multiple disturbance virtual event sets.

[0103] In the embodiment of the present application, the data processing device determines the sub-prediction result set corresponding to each disturbance virtual event set based on the sub-prediction result corresponding to each virtual event in the prediction result acquisition process, and fuses each sub-prediction result in the sub-prediction result set to obtain the label corresponding to each disturbance virtual event set, thereby obtaining multiple disturbance virtual event sets and labels corresponding to multiple disturbance virtual event sets. Then, the data processing device trains a fitting model based on multiple disturbance virtual event sets and labels corresponding to multiple disturbance virtual event sets.

[0104] It should be noted that the fitting model is obtained by simulating the process of obtaining the prediction results locally based on the target virtual event.

[0105] Step 4036: Based on the fitting model, determine the event contribution of the target virtual event to the prediction result.

[0106] In the embodiment of the present application, after the data processing device obtains the fitting model, it uses the fitting model to predict the result of the target virtual event, and the obtained result is the event contribution of the target virtual event to the predicted result.

[0107] It can be understood that the data processing device starts from the local part, obtains the fitting model by simulating the process of obtaining the prediction results, and obtains the event contribution of the target virtual event to the prediction results through the fitting model, so that the prediction results are explainable.

[0108] In the embodiment of the present application, the data processing device in step 403 determines the prediction result based on at least two virtual events, including steps 4037 to 4039 (not shown in the figure), and each step is described below.

[0109] Step 4037: extract features from each virtual event to obtain virtual event features, and determine sub-prediction results based on the virtual event features.

[0110] In the embodiment of the present application, when the image processing device determines the prediction result based on at least two virtual events, it is obtained based on the sub-prediction result corresponding to each virtual event; wherein the sub-prediction result corresponding to each virtual event is determined by the virtual event feature corresponding to the virtual event. Here, the sub-prediction result refers to the result obtained by performing the target prediction task based on the virtual event.

[0111] Step 4038: Obtain the target time weight corresponding to each virtual event from the specified time weight corresponding to each specified event type.

[0112] In an embodiment of the present application, a designated time weight corresponding to each designated event type is provided in the image processing device, or the image processing device can obtain the designated time weight corresponding to each designated event type from other devices, and the designated time weight represents the weight corresponding to each designated event type at different times, that is, the weight corresponding to each designated event type at different times is different, and the designated time weight is the correspondence between the time of occurrence of the event and the degree of importance, including multiple weights corresponding to multiple times. Here, the image processing device determines the weight matching the occurrence time from the designated event weight based on the occurrence time of each virtual event, and obtains the target time weight. In addition, at the same time, the cumulative sum of the weights corresponding to each designated event type is a specified value (for example, 1).

[0113] It should be noted that the target event weight is a weight set in advance for the same virtual scene application (application used to render virtual scenes) and is static; while the event contribution changes dynamically with different virtual scenes and is dynamic. Therefore, through attribution, dynamic quantification of prediction results can be achieved.

[0114] Step 4039: determine the fusion result of the target time weight and the sub-prediction result as the prediction result to be integrated, obtain at least two prediction results to be integrated corresponding to at least two virtual events, and integrate the at least two prediction results to be integrated to obtain a prediction result.

[0115] In the embodiment of the present application, the image processing device fuses the target time weight with the sub-prediction result, and determines the obtained fusion result as the prediction result to be integrated; thus, for at least two virtual events, at least two prediction results to be integrated can be obtained, wherein the at least two virtual events correspond to the at least two prediction results to be integrated one by one. Here, the image processing device can complete the integration by superimposing at least two prediction results to be integrated to obtain the prediction result.

[0116] In the embodiment of the present application, in step 4037, the image processing device extracts features of each virtual event to obtain virtual event features, including step 40371 and step 40372 (not shown in the figure), and each step is described below.

[0117] Step 40371: Based on the specified feature pool, determine the event change information corresponding to each virtual event.

[0118] It should be noted that a designated feature pool is provided in the image processing device, or the image processing device can obtain a designated feature pool from other devices, and the designated feature pool includes historical virtual events corresponding to each historical time; here, the image processing device determines the event change information corresponding to each current virtual event based on the historical virtual events in the designated feature pool, such as the growth of the virtual economy, the growth of the virtual kill volume, etc.

[0119] Step 40372: Extract features of virtual events and event change information to obtain virtual event features.

[0120] In the embodiment of the present application, the image processing device determines the features commonly corresponding to the virtual events and the event change information as virtual event features corresponding to each virtual event to complete feature extraction.

[0121] Accordingly, in step 402 of the embodiment of the present application, after the image processing device determines at least two virtual events corresponding to each specified event type based on the virtual activity data, the data processing method further includes: the image processing device stores each virtual event and the occurrence time corresponding to the virtual event in the specified feature pool to complete the update of the specified feature pool; thereby, the feature extraction of the virtual event at the next time is performed based on the updated specified feature pool.

[0122] It can be understood that by combining virtual events and event change information corresponding to virtual events, feature extraction is determined based on current virtual data and the change amount corresponding to the current virtual data, thereby improving the richness of virtual event features. Further, when performing target prediction tasks based on virtual event features, the accuracy of prediction results can be improved.

[0123] In the embodiment of the present application, steps 405 to 407 (not shown in the figure) are also included before step 403; that is, before the image processing device determines the prediction result based on at least two virtual events, the data processing method also includes steps 405 to 407, and each step is explained below.

[0124] Step 405: Determine a virtual event that is numerical data from at least two virtual events, and obtain a virtual event to be processed.

[0125] In an embodiment of the present application, the image processing device selects a virtual event belonging to numerical data from at least two virtual events, and determines the selected virtual event belonging to numerical data as a virtual event to be processed; thus, the virtual event to be processed includes at least one virtual event belonging to numerical data.

[0126] Step 406: normalize the virtual event to be processed to obtain the virtual event to be replaced.

[0127] In the embodiment of the present application, the image processing device normalizes each virtual event in the virtual events to be processed to obtain the normalized virtual events to be processed, and the normalized virtual events to be processed are the virtual events to be replaced.

[0128] Step 407: Based on the virtual event to be replaced, replace the corresponding virtual event in at least two virtual events to obtain at least two first virtual events to be predicted.

[0129] It should be noted that the image processing device updates the virtual event to be replaced to at least two virtual events, thereby obtaining at least two first virtual events to be predicted; wherein the updating process is to replace the corresponding virtual event in at least two virtual events with the virtual event to be replaced.

[0130] Accordingly, in the embodiment of the present application, step 403, in which the image processing device determines a prediction result based on at least two virtual events, includes: the image processing device determines a prediction result based on at least two first virtual events to be predicted.

[0131] Before step 403, steps 408 to 410 (not shown in the figure) are also included; that is, before the image processing device determines the prediction result based on at least two virtual events, the data processing method also includes steps 408 to 410, and each step is explained below.

[0132] Step 408: Determine a virtual event belonging to category data from at least two virtual events to obtain a virtual event to be converted.

[0133] In an embodiment of the present application, the image processing device selects a virtual event belonging to category type data from at least two virtual events, and determines the selected virtual event belonging to category type data as the virtual event to be converted; thus, the conversion event to be processed includes at least one virtual event belonging to category type data.

[0134] Step 409: Obtain vector information corresponding to the virtual event to be converted, and obtain a virtual event vector.

[0135] Step 410: Based on the virtual event vector, replace the corresponding virtual event in at least two virtual events to obtain at least two second virtual events to be predicted.

[0136] Accordingly, in the embodiment of the present application, step 403, in which the image processing device determines a prediction result based on at least two virtual events, includes: the image processing device determines a prediction result based on at least two second virtual events to be predicted.

[0137] In the embodiment of the present application, the prediction result can be obtained through a result prediction model; wherein the result prediction model can be obtained through steps 411 to 413 (not shown in the figure), and each step is described below.

[0138] Step 411: Determine the result prediction model to be trained based on the prediction result type.

[0139] In the embodiment of the present application, the prediction result type includes at least one of a regression type and a classification type; wherein, when the target prediction task is a regression task, the image processing device determines the prediction result type as a regression type, and when the target prediction task is a classification type, the image processing device determines the prediction result type as a classification type. wherein the result prediction model to be trained is a network model to be trained for performing the target prediction task.

[0140] Step 412: Based on the prediction model for the result to be trained, predict the feature training samples to obtain an estimated prediction result.

[0141] In the embodiment of the present application, the image processing device is provided with feature training samples, or the image processing device can obtain feature training samples from other devices, and the feature training samples are feature samples of each virtual event, which are used to train the prediction model for the result to be trained. Thus, the image processing device uses the prediction model for the result to be trained to perform the target prediction task on the feature training samples, and obtains the estimated prediction result.

[0142] Step 413: Based on the loss between the estimated prediction result and the prediction result label, the result prediction model to be trained is trained to obtain a result prediction model.

[0143] It should be noted that the image processing device is provided with a prediction result label, or the image processing device can obtain the prediction result label from other devices, and the prediction result label is the label of the feature training sample. Therefore, the image processing device performs back propagation in the result prediction model to be trained based on the loss between the estimated prediction result and the prediction result label to adjust the parameters in the result prediction model to be trained, realizes the training of the result prediction model to be trained, and thus trains the result prediction model. Among them, the result prediction model is the trained result prediction model to be trained.

[0144] In the embodiment of the present application, the event contribution can be obtained through a contribution prediction model; wherein the contribution prediction model can be obtained through steps 414 and 415 (not shown in the figure), and each step is described below.

[0145] Step 414: Based on the contribution prediction model to be trained, the contribution training samples are predicted to obtain an estimated contribution.

[0146] In the embodiment of the present application, the image processing device is provided with contribution training samples, or the image processing device can obtain contribution training samples from other devices, and the contribution training samples are the influence of each virtual event sample on the predicted result label. Thus, the image processing device attributes the contribution training samples using the contribution prediction model to be trained, and obtains the estimated contribution.

[0147] Step 415: Based on the loss between the estimated contribution and the contribution label, the contribution prediction model to be trained is trained to obtain a contribution prediction model.

[0148] It should be noted that the image processing device is provided with a contribution label, or the image processing device can obtain a contribution label from other devices, and the contribution label is a label of a contribution training sample. Therefore, the image processing device performs back propagation in the contribution prediction model to be trained based on the loss between the estimated contribution and the contribution label to adjust the parameters in the contribution prediction model to be trained, thereby training the contribution prediction model to be trained, and thus training a contribution prediction model. Among them, the contribution prediction model is a trained contribution prediction model to be trained.

[0149] In an embodiment of the present application, after the image processing device displays the prediction result and at least two event contributions corresponding to at least two virtual events in step 404, the data processing method further includes: when the update time arrives, the image processing device obtains new virtual activity data, wherein the update time is determined based on the update cycle; based on at least two new virtual events determined by the new virtual activity data, a new prediction result is determined, and a new event contribution of each new virtual event to the new prediction result is determined, and at least two new event contributions corresponding to at least two new virtual events are obtained; the new prediction result and at least two new event contributions corresponding to at least two new virtual events are displayed. In this way, real-time attribution can be achieved.

[0150] In an embodiment of the present application, after the image processing device displays the prediction results and at least two event contributions corresponding to at least two virtual events in step 404, the data processing method also includes: the image processing device performs a target interaction operation in the virtual scene in response to the interaction operation received for the displayed prediction results and at least two event contributions corresponding to at least two virtual events, wherein the target interaction operation is an interaction operation associated with a virtual event with a contribution less than a specified degree.

[0151] Below, an exemplary application of the embodiment of the present application in an actual application scenario will be described. The exemplary application describes a process in which the data processing method provided by the embodiment of the present application can automatically obtain the explanation content (called the prediction result, and at least two event contributions corresponding to at least two virtual events) of the game scene (called the virtual scene) in the artificial intelligence explanation system.

[0152] See also Figure 8 , Figure 8 is a schematic diagram of an exemplary data processing method provided in an embodiment of the present application; Figure 8 As shown, the exemplary data processing method includes a feature extraction stage 8-1, a target analysis stage 8-2 and an event attribution stage 8-3.

[0153] In the feature extraction stage 8-1, the real-time game data (called virtual activity data) in the current game scene 8-4 is acquired, and based on different pre-set fields (called designated event types), the real-time game data is divided into corresponding game events (called at least two virtual events), and the game events in each field are represented in the form of feature groups; then, the feature group of the game events in each field is projected into the corresponding vector space, and the spatial model corresponding to the game events in each field is used to extract the key information in the projection result, so as to obtain the subsequent event features to be processed corresponding to the game events in each field (called virtual event features).

[0154] Here, when the current game is a multiplayer online tactical competitive game (Multiplayer Online Battle Arena, MOBA), the pre-set different fields may include a virtual economy field, a virtual kill field, a virtual defense resource field, a virtual wild area resource field, a virtual soldier line field, and a virtual lineup field. Among them, the virtual economy field corresponds to information such as the virtual economic difference accumulated by the two teams at a certain moment, which is positively correlated with the predicted winning rate; the virtual kill field corresponds to information such as the difference in the number of virtual kills of the two teams, which is positively correlated with the predicted winning rate. The virtual defense resource field corresponds to information such as the number of virtual defense resources, which is positively correlated with the predicted winning rate. The virtual soldier line field corresponds to information such as the distance from the virtual soldier line to the target virtual land of the virtual enemy. The virtual lineup field corresponds to information such as the selected virtual object, and different virtual objects correspond to different attributes and virtual skills, and different virtual objects correspond to different relationships, such as cooperation relationships, restraint relationships, etc. Therefore, when the real-time game data in the current game scene 8-4 is divided, the following can be obtained: Figure 8 The division results are 8-41.

[0155] In addition, when the current game is a first-person shooting game (FPSG), the pre-set different fields may include a virtual kill field, a virtual field of vision, a virtual sound field, a virtual position field, a virtual equipment field, and a virtual scene field. Among them, the virtual kill field corresponds to information such as the difference in the number of virtual kills of the two teams. The virtual field of vision corresponds to information such as the range and corresponding content of the virtual field of vision. The virtual sound field corresponds to information such as the acquired sound, such as footsteps, car engine sounds, and explosive sounds, which are positively correlated with the predicted winning rate. The virtual position field corresponds to information such as the distance of the virtual object, and the relative position with virtual teammates and virtual enemies. The virtual equipment field corresponds to information such as virtual weapons, virtual armor, virtual ammunition, and virtual medical supplies equipped by the virtual object. The virtual scene field corresponds to information such as relevant information of the virtual scene, such as a virtual damage circle.

[0156] It should also be noted that different feature groups will undergo different forms of preprocessing steps before feature extraction. The three corresponding game events in the virtual economy field, virtual kill field, and virtual defense resource field are numerical data, and the corresponding preprocessing steps are normalization based on the maximum and minimum values; the three corresponding game events in the virtual jungle resource field, virtual soldier line field, and virtual lineup field are categorical data, and the corresponding preprocessing steps are conversion into vector form.

[0157] In the target analysis stage 8-2, based on the event features obtained in the feature extraction stage 8-1, the prediction results (90% winning rate of the red team) and related game events are determined. Among them, the prediction task is determined to be a classification task or a regression task based on the prediction target; for example, when the prediction target is to determine the winning virtual team, or to determine the virtual object that completes the specified virtual operation, the prediction task is determined to be a classification task; when the prediction target is to determine the virtual economy after 5 minutes from the current time, or to determine the virtual game score after 3 minutes, the prediction task is determined to be a regression task. Here, for classification tasks, various classification models can be used, such as random forest models, additive models (eXtreme Gradient Boosting, XGBoost), deep neural network models, decision tree models and logistic regression models; for regression tasks, various regression models can be used, such as discrete regression models, linear regression models, random forest regression tree models and deep neural network models.

[0158] It should be noted that a corresponding sub-prediction result can be obtained based on each event feature. By obtaining the importance weight corresponding to the current time (called the target time weight) and calculating the importance weight multiplied by the sub-prediction result, the prediction part corresponding to each event feature group for determining the current prediction result is obtained; finally, by superimposing the prediction parts corresponding to each event feature group, the current prediction result (for example, winning rate, score, etc.) is obtained.

[0159] In the event attribution stage 8-3, the prediction results obtained in the target analysis stage 8-2 are further processed to obtain the contribution of each game event to the prediction results (the contribution of the red team's virtual economy leading by 2k is 0.3, the contribution of the red team's virtual kill count leading by 5 is 0.65, and the contribution of the blue team's late virtual lineup advantage is -0.05, which is called event contribution) to determine the degree of influence of each game event on the prediction results.

[0160] Here, intrinsic interpretable machine learning can be used for direct learning to obtain the contribution of each game event to the prediction result, and the contribution of each game event to the prediction result can also be obtained by post-attribution (for example, local interpretable model (LIME), Shapley value model, etc.). When the Shapley value model is used for post-attribution, the contribution of each game event i to the prediction result can be obtained by formula (1), which is as follows.

[0161]

[0162] Among them, v() is used to determine the contribution corresponding to the subset, N refers to the number of game events, and S is the subset of game event i before i in N; is the Shapley value of game event i, that is, the contribution to the prediction result; S c N\{i} means that S is a subset of N that does not include game event i, v(S∪{i})-v(S) represents the contribution of game event i to S (called sequence contribution), Represents the probability of game event i joining S.

[0163] It should be noted that the Shapley value is mainly used to solve the problem of distribution equilibrium in cooperative game theory. The data set usually contains many feature variables (corresponding to game events in the game). The prediction results obtained using the characteristics of the data set can be regarded as the completion of the game events in cooperation. The Shapley value is the contribution of each game event to the prediction result.

[0164] It should also be noted that the target analysis stage 8-2 is used to accurately obtain the prediction results. The event attribution stage 8-3 is used to quantify the contribution of game events; wherein, the contribution is a time-varying value, because the importance of various game events occurring in the virtual game is different at different times; at the same time, the contribution can be a positive contribution or a negative contribution. When the contribution is positive, the occurrence of the corresponding game event has an incentive effect on the current prediction result (called a positive effect); otherwise, it has an inhibitory effect (called a negative effect).

[0165] In the embodiment of the present application, the three stages are processed separately, that is, the three stages are processed independently and are decoupled from each other.

[0166] For example, see Fig. 9 , Fig. 9 is an exemplary data processing schematic diagram provided in the embodiment of the present application; Fig. 9As shown, page 9-1 displays a virtual scene 9-11 and a button 9-12 corresponding to the virtual scene; when an operation of clicking button 9-12 is received, the real-time game data corresponding to the virtual scene 9-11 is obtained for data processing, and the data processing result 9-13 obtained by the data processing is displayed. Among them, the prediction result (68% winning rate) is based on Figure 8 The 6 pieces of information that affect the prediction results are based on the target analysis stage 8-2. Figure 8 First, based on the 6 pieces of information that affect the prediction results in the data processing results 9-13, it can be determined that there are 4 game events that are conducive to the blue team's victory, namely "virtual lineup late advantage", "virtual attack lead", "taking designated resources" and "virtual damage output high", which are represented by pattern-filled bars; and each game event also corresponds to a quantitative contribution, namely the game event "virtual lineup late advantage" corresponding to the contribution is 39%, that is, 39% of the predicted 68% winning rate comes from this game event, that is, this game event is the most important basis for judging the final victory of the blue team. On the other hand, in addition to positive game events, there are also negative game events that can reduce the predicted results, such as "virtual economy lags behind 2k", which leads to a 3% drop in the winning rate.

[0167] It should be noted that, for the overall game level, the target analysis stage 8-2 corresponds to the global event weight, which is used to obtain the impact of each game event on the overall prediction result (such as win rate or economy) at different occurrence times. For a single game, attribution can provide dynamic event impact analysis and can perform dynamic analysis.

[0168] It is understandable that using attribution to determine the impact of game events on target indicators (such as wins, losses, competition or scores, etc.) can not only output the real-time winning rate of both teams based on the real-time data of the game, but also provide quantitative attribution information for game events, dynamically process data on events, refine the granularity of target-oriented data processing, and improve interpretability.

[0169] The following further describes an exemplary structure of the data processing device 255 provided in the embodiment of the present application implemented as a software module. In some embodiments, Figure 2 As shown, the software modules stored in the data processing device 255 of the memory 250 may include:

[0170] A data acquisition module 2551 is used to display a virtual scene and acquire virtual activity data in response to a data processing operation on the virtual scene, wherein the virtual activity data is data associated with an interactive operation in the virtual scene;

[0171] An event determination module 2552, configured to determine at least two virtual events corresponding to respective specified event types based on the virtual activity data, wherein the specified event types are virtual event types in the virtual scene;

[0172] The data processing module 2553 is used to determine the prediction result based on the at least two virtual events, and determine the event contribution of each virtual event to the prediction result based on the at least two virtual events, so as to obtain at least two event contributions corresponding to the at least two virtual events;

[0173] The result display module 2554 is used to display the prediction result and at least two event contributions corresponding to at least two virtual events.

[0174] In an embodiment of the present application, the data processing module 2553 is also used to determine multiple virtual event sequences including a target virtual event based on at least two of the virtual events, wherein the target virtual event is any one of the at least two virtual events; obtain the sequence contribution of the target virtual event in each of the virtual event sequences, and obtain multiple sequence contributions corresponding to the multiple virtual event sequences; and determine the average of the multiple sequence contributions as the event contribution of the target virtual event to the prediction result.

[0175] In an embodiment of the present application, the data processing module 2553 is also used to perturb at least two of the virtual events based on a target virtual event among the at least two virtual events to obtain multiple sets of perturbed virtual events; train a fitting model based on the multiple sets of perturbed virtual events; and determine the event contribution of the target virtual event to the prediction result based on the fitting model.

[0176] In an embodiment of the present application, the data processing module 2553 is also used to extract features from each of the virtual events to obtain virtual event features; determine sub-prediction results based on the virtual event features; obtain the target time weight corresponding to each of the virtual events from the specified time weight corresponding to each of the specified event types, wherein the specified time weight is the correspondence between the event occurrence time and the importance; determine the fusion result of the target time weight and the sub-prediction result as the prediction result to be integrated, and obtain at least two prediction results to be integrated corresponding to at least two of the virtual events; integrate at least two of the prediction results to be integrated to obtain the prediction result.

[0177] In an embodiment of the present application, the data processing module 2553 is also used to determine the event change information corresponding to each of the virtual events based on a specified feature pool, wherein the specified feature pool includes historical virtual events corresponding to each historical time; and perform feature extraction on the virtual events and the event change information to obtain the virtual event features.

[0178] In the embodiment of the present application, the data processing module 2553 is further used to store each of the virtual events and the time corresponding to the virtual events in the designated feature pool.

[0179] In an embodiment of the present application, the data processing module 2553 is also used to determine the virtual event that belongs to numerical data from at least two of the virtual events to obtain a virtual event to be processed; normalize the virtual event to be processed to obtain a virtual event to be replaced; and replace the corresponding virtual event in at least two of the virtual events based on the virtual event to be replaced to obtain at least two first virtual events to be predicted.

[0180] In the embodiment of the present application, the data processing module 2553 is further used to determine the prediction result based on at least two of the first virtual events to be predicted.

[0181] In an embodiment of the present application, the data processing module 2553 is also used to determine the virtual event belonging to category data from at least two of the virtual events to obtain the virtual event to be converted; obtain the vector information corresponding to the virtual event to be converted to obtain a virtual event vector; based on the virtual event vector, replace the corresponding virtual event in at least two of the virtual events to obtain at least two second virtual events to be predicted.

[0182] In the embodiment of the present application, the data processing module 2553 is further used to determine the prediction result based on at least two of the second virtual events to be predicted.

[0183] In an embodiment of the present application, the prediction result is obtained through a result prediction model, and the data processing device 255 also includes a model training module 2555, which is used to determine the result prediction model to be trained based on the prediction result type, wherein the prediction result type includes at least one of a regression type and a classification type; based on the result prediction model to be trained, the feature training sample is predicted to obtain an estimated prediction result; based on the loss between the estimated prediction result and the prediction result label, the result prediction model to be trained is trained to obtain the result prediction model, wherein the prediction result label is the label of the feature training sample.

[0184] In an embodiment of the present application, the event contribution is obtained through a contribution prediction model, and the model training module 2555 is also used to predict the contribution training sample based on the contribution prediction model to be trained to obtain an estimated contribution; based on the loss between the estimated contribution and the contribution label, the contribution prediction model to be trained is trained to obtain the contribution prediction model, wherein the contribution label is the label of the contribution training sample.

[0185] In an embodiment of the present application, the data processing module 2553 is also used to obtain new virtual activity data when the update time arrives, wherein the update time is determined based on the update cycle; determine a new prediction result based on at least two new virtual events determined by the new virtual activity data, and determine a new event contribution of each new virtual event to the new prediction result, and obtain at least two new event contributions corresponding to at least two new virtual events; display the new prediction result and at least two new event contributions corresponding to at least two new virtual events.

[0186] In an embodiment of the present application, the data processing module 2553 is also used to perform a target interaction operation in the virtual scene in response to the interactive operations received for the displayed prediction results and at least two event contributions corresponding to at least two of the virtual events, wherein the target interaction operation is an interactive operation associated with the virtual event with a contribution less than a specified degree.

[0187] The embodiment of the present application provides a computer program product or a computer program, which includes computer instructions stored in a computer-readable storage medium. The processor of a computer device (referred to as a data processing device) reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device executes the data processing method described above in the embodiment of the present application.

[0188] The present application embodiment provides a computer-readable storage medium storing executable instructions, wherein the executable instructions are stored. When the executable instructions are executed by a processor, the processor will execute the data processing method provided by the present application embodiment, for example, Figure 4 The data processing method is shown.

[0189] In some embodiments, the computer-readable storage medium may be a memory such as FRAM, ROM, PROM, EP ROM, EEPROM, flash memory, magnetic surface storage, optical disk, or CD-ROM; or it may be various devices including one or any combination of the above memories.

[0190] In some embodiments, executable instructions may be in the form of a program, software, software module, script or code, written in any form of programming language (including compiled or interpreted languages, or declarative or procedural languages), and may be deployed in any form, including as a stand-alone program or as a module, component, subroutine or other unit suitable for use in a computing environment.

[0191] As an example, executable instructions may, but need not, correspond to a file in a file system, may be stored as part of a file that stores other programs or data, such as in one or more scripts in a HyperText Markup Language (HTML) document, in a single file dedicated to the program in question, or in multiple coordinated files (e.g., files storing one or more modules, subroutines, or code portions).

[0192] As an example, executable instructions may be deployed to be executed on one computer device (in which case, the one computer device is a data processing device), or on multiple computer devices located at one location (in which case, the multiple computer devices located at one location are data processing devices), or on multiple computer devices distributed at multiple locations and interconnected by a communication network (in which case, the multiple computer devices distributed at multiple locations and interconnected by a communication network are data processing devices).

[0193] It is understandable that in the embodiments of the present application, when virtual activity data and other related data are involved, when the embodiments of the present application are applied to specific products or technologies, user permission or consent is required, and the collection, use and processing of relevant data need to comply with relevant laws, regulations and standards of relevant countries and regions.

[0194] In summary, the embodiment of the present application determines at least two virtual events by collecting virtual activity data in a virtual scene, determines a prediction result based on the at least two virtual events, and attributes the event contribution of each relevant virtual event to the prediction result; so that not only at least two virtual events associated with the prediction result can be obtained, but also the degree of influence of each virtual event on the prediction result can be obtained; thereby, the granularity of the processing results can be refined, the comprehensiveness of the data display can be improved, and the effect of data processing can be improved.

[0195] The above is only an embodiment of the present application and is not intended to limit the protection scope of the present application. Any modifications, equivalent substitutions and improvements made within the spirit and scope of the present application are included in the protection scope of the present application.

Claims

1. A data processing method, characterized in that: The method comprises: Displaying a virtual scene, and acquiring virtual activity data in response to a data processing operation on the virtual scene, wherein the virtual activity data is data associated with an interactive operation in the virtual scene; Based on the virtual activity data, determining at least two virtual events corresponding to respective designated event types, wherein the designated event types are virtual event types in the virtual scene; Determine a prediction result based on at least two of the virtual events, and determine an event contribution of each of the virtual events to the prediction result based on the at least two virtual events, to obtain at least two event contributions corresponding to the at least two virtual events; Displaying the prediction result and at least two event contributions corresponding to at least two virtual events; The method of determining a prediction result based on at least two of the virtual events includes: determining event change information corresponding to each of the virtual events based on a designated feature pool, wherein the designated feature pool includes historical virtual events corresponding to each historical time; performing feature extraction on the virtual events and the event change information to obtain virtual event features; determining sub-prediction results based on the virtual event features; obtaining a target time weight corresponding to each of the virtual events from a designated time weight corresponding to each of the designated event types, wherein the designated time weight is a correspondence between an event occurrence time and its importance; determining a fusion result of the target time weight and the sub-prediction results as a prediction result to be integrated, and obtaining at least two prediction results to be integrated corresponding to at least two of the virtual events; and integrating at least two of the prediction results to be integrated to obtain the prediction result.

2. The method according to claim 1, characterized in that: The determining, based on at least two of the virtual events, the event contribution of each of the virtual events to the prediction result comprises: Based on at least two of the virtual events, determining a plurality of virtual event sequences including a target virtual event, wherein the target virtual event is any one of the at least two virtual events; Acquire the sequence contribution of the target virtual event in each of the virtual event sequences, and obtain a plurality of sequence contributions corresponding to the plurality of virtual event sequences; The average of the contribution degrees of the multiple sequences is determined as the event contribution degree of the target virtual event to the prediction result.

3. The method according to claim 1, characterized in that The determining, based on at least two of the virtual events, the event contribution of each of the virtual events to the prediction result comprises: Based on a target virtual event among at least two of the virtual events, perturbing at least two of the virtual events to obtain a plurality of perturbed virtual event sets; Based on the plurality of disturbance virtual event sets, training a fitting model; Based on the fitting model, the event contribution of the target virtual event to the prediction result is determined.

4. The method according to claim 1, characterized in that After determining at least two virtual events corresponding to each specified event type based on the virtual activity data, the method further includes: Each of the virtual events and the time corresponding to the virtual events are stored in the designated feature pool.

5. The method according to any one of claims 1 to 3, characterized in that: Before determining the prediction result based on at least two of the virtual events, the method further includes: Determine the virtual event that is numerical data from at least two of the virtual events to obtain a virtual event to be processed; Normalizing the virtual event to be processed to obtain a virtual event to be replaced; Based on the virtual event to be replaced, replace the corresponding virtual event in at least two virtual events to obtain at least two first virtual events to be predicted; The step of determining a prediction result based on at least two of the virtual events comprises: The prediction result is determined based on at least two of the first virtual events to be predicted.

6. The method according to any one of claims 1 to 3, characterized in that: Before determining the prediction result based on at least two of the virtual events, the method further includes: Determine the virtual event belonging to category data from at least two of the virtual events to obtain a virtual event to be converted; Acquire vector information corresponding to the virtual event to be converted to obtain a virtual event vector; Based on the virtual event vector, replace the corresponding virtual events in at least two of the virtual events to obtain at least two second virtual events to be predicted; The step of determining a prediction result based on at least two of the virtual events comprises: The prediction result is determined based on at least two of the second virtual events to be predicted.

7. The method according to any one of claims 1 to 3, characterized in that: The prediction result is obtained by a result prediction model, and the result prediction model is obtained by the following steps: Determining a result prediction model to be trained based on the prediction result type, wherein the prediction result type includes at least one of a regression type and a classification type; Based on the prediction model for the result to be trained, predict the feature training samples to obtain an estimated prediction result; Based on the loss between the estimated prediction result and the prediction result label, the result prediction model to be trained is trained to obtain the result prediction model, wherein the prediction result label is the label of the feature training sample.

8. The method according to any one of claims 1 to 3, characterized in that: The event contribution is obtained through a contribution prediction model, and the contribution prediction model is obtained through the following steps: Based on the contribution prediction model to be trained, the contribution training samples are predicted to obtain the estimated contribution; Based on the loss between the estimated contribution and the contribution label, the contribution prediction model to be trained is trained to obtain the contribution prediction model, wherein the contribution label is a label of the contribution training sample.

9. The method according to any one of claims 1 to 3, characterized in that: After displaying the prediction result and at least two event contributions corresponding to at least two virtual events, the method further includes: When an update time arrives, new virtual activity data is acquired, wherein the update time is determined based on an update cycle; Determine a new prediction result based on at least two new virtual events determined by the new virtual activity data, and determine a new event contribution of each of the new virtual events to the new prediction result, and obtain at least two new event contributions corresponding to the at least two new virtual events; The new prediction result and at least two new event contributions corresponding to at least two new virtual events are displayed.

10. The method according to any one of claims 1 to 3, characterized in that: After displaying the prediction result and at least two event contributions corresponding to at least two virtual events, the method further includes: In response to the interactive operations received for the displayed prediction results and at least two event contributions corresponding to at least two of the virtual events, a target interactive operation is performed in the virtual scene, wherein the target interactive operation is an interactive operation associated with the virtual event with a contribution less than a specified degree.

11. A data processing device, characterized in that: The data processing device comprises: A data acquisition module, used to display a virtual scene, and acquire virtual activity data in response to a data processing operation on the virtual scene, wherein the virtual activity data is data associated with an interactive operation in the virtual scene; An event determination module, configured to determine at least two virtual events corresponding to respective specified event types based on the virtual activity data, wherein the specified event types are virtual event types in the virtual scene; A data processing module, used to determine a prediction result based on at least two of the virtual events, and determine an event contribution of each of the virtual events to the prediction result based on the at least two virtual events, to obtain at least two event contributions corresponding to the at least two virtual events; A result display module, used for displaying the prediction result and at least two event contributions corresponding to at least two virtual events; The data processing module is also used to determine the event change information corresponding to each of the virtual events based on a specified feature pool, wherein the specified feature pool includes historical virtual events corresponding to each historical time; perform feature extraction on the virtual events and the event change information to obtain virtual event features; determine sub-prediction results based on the virtual event features; obtain the target time weight corresponding to each of the virtual events from the specified time weight corresponding to each of the specified event types, wherein the specified time weight is the correspondence between the event occurrence time and the importance; determine the fusion result of the target time weight and the sub-prediction results as the prediction result to be integrated, and obtain at least two prediction results to be integrated corresponding to at least two of the virtual events; integrate at least two of the prediction results to be integrated to obtain the prediction result.

12. A data processing device, characterized in that: The data processing device comprises: A memory for storing executable instructions; A processor, configured to implement the data processing method according to any one of claims 1 to 10 when executing the executable instructions stored in the memory.

13. A computer-readable storage medium storing executable instructions, characterized in that: The executable instructions are used to implement the data processing method according to any one of claims 1 to 10 when executed by a processor.

14. A computer program product comprising a computer program or instructions, characterized in that When the computer program or instruction is executed by a processor, the data processing method according to any one of claims 1 to 10 is implemented.

Images