Virtual commodity price prediction method and device, equipment and storage medium

Abstract

The invention provides a virtual commodity price prediction method and device, equipment and a storage medium, and the method comprises the steps: determining a target channel as a monitoring target in preset message push channels based on the type of an operation system of a user terminal; under the condition that a plurality of currently published system messages are monitored from the target channel, determining a target message associated with the virtual commodity in the plurality of system messages, and analyzing the target message to obtain corresponding text feature information; and determining a target virtual commodity and a price change trend of the target virtual commodity according to the text feature information, and determining a predicted price of the target virtual commodity based on the price change trend. According to the scheme, the virtual commodity price can be predicted in combination with the system message, so that accurate and reliable reference data is provided for the user.

Description

Technical Field

[0001] This application relates to the field of computer technology, and in particular to a method, apparatus, device and storage medium for predicting virtual commodity prices. Background Technology

[0002] Currently, some game applications, business applications, and other software applications that can simulate transactions provide users with the function of trading virtual goods. Through the built-in trading platform, users can sell the virtual goods they obtain in the software application to obtain virtual currency used in the software application, or use virtual currency to purchase virtual goods on the trading platform, thereby increasing users' sense of participation and user experience, so as to further increase the number of active users.

[0003] However, in practical applications, while trading platforms push messages that trigger price fluctuations for virtual goods, the price adjustment mechanisms in related technologies are flawed. For example, some software applications use the lowest transaction price over several recent trading periods as the overall price of the virtual goods on the platform, while others allow operators to customize the pricing, resulting in significant price differences across different trading periods. This inconsistency in price adjustment mechanisms among different software applications leads to insufficient information, preventing users from making informed decisions about buying or selling virtual goods and profiting from them. Even within the same software application, price changes across different trading periods can be irregular due to unreasonable price adjustment mechanisms, failing to provide users with a reliable reference for buying or selling virtual goods. Therefore, it hinders users from making informed decisions based on virtual goods prices; in other words, there is a lack of a reasonable and efficient mechanism to address the technical problems of insufficient information accuracy and poor reference value caused by unreasonable virtual goods pricing in existing technologies. Summary of the Invention

[0004] This application provides a method, apparatus, server, and storage medium for predicting virtual commodity prices, which solves the technical problem in related technologies where the price adjustment mechanism of the trading platform is unreasonable, resulting in insufficient information accuracy and poor reference value. This solution can reasonably predict the price of virtual commodities and provide users with accurate and reliable reference data.

[0005] Firstly, this application provides a method for predicting the price of virtual goods, which includes: Select the target channel as the monitoring target from the preset message push channels; When multiple system messages are detected from the target channel, the target message associated with the virtual product is identified from among the multiple system messages, and the target message is parsed to obtain the corresponding text feature information. Based on textual features, the target virtual product and its price trend are determined. Based on the price trend, the predicted price of the target virtual product is determined.

[0006] Secondly, this application also provides a virtual commodity price prediction device, which includes: The channel selection module is configured to determine the target channel as the monitoring target among preset message push channels; The message parsing module is configured to, when listening to multiple currently published system messages from the target channel, identify the target message associated with the virtual product among the multiple system messages, and parse the target message to obtain the corresponding text feature information; The price prediction module is configured to determine the target virtual product and its price change trend based on text feature information, and then determine the predicted price of the target virtual product based on the price change trend.

[0007] Thirdly, this application also provides an electronic device comprising: One or more processors; A storage device for storing one or more programs, which, when executed by one or more processors, enable the one or more processors to implement the virtual commodity price prediction method of this application.

[0008] Fourthly, this application also provides a storage medium for storing computer-executable instructions, which, when executed by a processor, are used to execute the virtual commodity price prediction method of this application.

[0009] This application's solution listens to relevant message push channels and receives system messages, then extracts textual feature information from them. This textual feature information is used to determine the price change trend of the corresponding target virtual goods, thus combining system messages to judge virtual goods price changes. This makes the predicted prices provided to users more reliable. Furthermore, predicting virtual goods prices based on price change trends can combine the correlation between virtual goods and price changes, thereby improving the accuracy of virtual goods price predictions and facilitating the generation of accurate and reliable predicted prices as reference data. Attached Figure Description

[0010] Figure 1 A schematic diagram illustrating the steps of a virtual goods price prediction method provided in an embodiment of this application.

[0011] Figure 2 This is a schematic diagram illustrating the steps for determining a target virtual commodity and its price change trend, provided in an embodiment of this application.

[0012] Figure 3This is a schematic diagram illustrating the steps for determining a price change trend according to an embodiment of this application.

[0013] Figure 4 A schematic diagram of the structure of a virtual commodity price prediction device provided in an embodiment of this application.

[0014] Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0015] The embodiments of this application will be further described in detail below with reference to the accompanying drawings and examples. The specific embodiments described herein are merely illustrative of the embodiments of this application and are not intended to limit the scope of this application. It should also be noted that, for ease of description, only the parts related to the embodiments of this application are shown in the accompanying drawings, not all structures. Those skilled in the art, after reading this specification, should be able to deduce that any combination of technical features can constitute an optional implementation method, provided that the technical features do not contradict each other.

[0016] The terms "first," "second," etc., used in the specification and claims of this application are used to distinguish similar objects, not to describe a specific order or sequence. It should be understood that such use of data can be interchanged where appropriate so that embodiments of this application can be implemented in orders other than those illustrated or described herein, and the objects distinguished by "first," "second," etc., are generally of the same class, not limited in number; for example, a first object can be one or more. Furthermore, in the specification and claims, "and / or" indicates at least one of the connected objects, and the character " / " generally indicates that the preceding and following objects are in an "or" relationship. In the description of this application, "multiple" means two or more, and "several" means one or more.

[0017] In some game and business applications that simulate transactions, users earn virtual currency by completing tasks, which they then use to purchase virtual goods within the application. Virtual currency is essentially an in-app currency, existing solely within the application, while virtual goods are tradable virtual items traded using virtual currency. The application also provides a built-in trading platform where users can sell their acquired virtual goods for virtual currency within the application, or use virtual currency to purchase virtual goods on the platform. This increases user engagement and experience, ultimately boosting the number of active users.

[0018] However, in practical applications, trading platforms push messages that cause changes in the price of virtual goods, such as increased production or improved manufacturing processes, leading to price fluctuations within the corresponding trading period. However, existing technologies have flawed price adjustment mechanisms for virtual goods. For example, some software applications use the lowest transaction price across multiple recent trading periods as the platform's overall pricing, while others allow operators to customize pricing, resulting in significant price differences across different trading periods. Consequently, users cannot make informed decisions about trading virtual goods to maximize profits, nor can they obtain relevant reference information. In other words, there is a lack of a reasonable and efficient mechanism to address the technical problems of insufficient information accuracy and poor reference value caused by unreasonable virtual goods pricing in existing technologies, thus hindering users from making informed decisions based on virtual goods prices.

[0019] In this regard, this application provides a method for predicting the price of virtual goods. This method can be applied to electronic devices, such as mobile phones, tablets and other user terminals, to predict the price of virtual goods in software applications that can simulate transactions, such as game applications and business applications, so as to enable users to formulate corresponding trading strategies. Figure 1 A schematic diagram illustrating the steps of a virtual goods price prediction method provided in an embodiment of this application is shown in the figure. This solution can predict the price of virtual goods by combining system messages provided within the application. The specific steps include: Step S110: Determine the target channel as the monitoring target in the preset message push channels.

[0020] The user terminal, as a device running the aforementioned game applications, business applications, and other software applications that simulate transactions, is equipped with a corresponding operating system. For example, the user terminal's operating system may be iOS or Android, with each operating system corresponding to a different operating system type. Within the application, system messages are messages released by the application service provider through the server to enhance the application's playability. These messages may include those related to user account rewards or changes in virtual goods. The application service provider then sends system messages to the user terminal through the server, causing the user terminal to display the corresponding system messages on its interface. Optionally, a message push channel is used to transmit messages between the user terminal and the server; that is, system messages are sent through the message push channel.

[0021] Optionally, the server sends system messages to the corresponding user terminals via a message push channel based on a message middleware. The server uses different message push channels for user terminals with different operating system types. To this end, the user terminal determines a target channel from the preset message push channels. This target channel is associated with the device's own operating system type, and the user terminal designates this target channel as the listening target, so that the user terminal can obtain system information by listening to this target channel.

[0022] Step S120: When multiple system messages are detected from the target channel, identify the target message associated with the virtual product among the multiple system messages, and parse the target message to obtain the corresponding text feature information.

[0023] Optionally, the server may send multiple system messages through the target channel. Upon receiving these multiple system messages, the user terminal needs to distinguish them, specifically identifying the target message associated with the virtual goods. Optionally, each system message may be configured with a corresponding message category and message tag to differentiate between them. For example, system messages can be categorized into three types: a first type associated with user accounts (such as changes in the amount of virtual currency, virtual goods, and other virtual resources held by the user account); a second type associated with changes in virtual goods on the trading platform; and a third type associated with task postings. These three types of messages correspond to different message categories, and system messages within the same category can be distinguished by different message tags. For instance, for the second type of message, different message tags can be used to set the associated virtual goods. For example, a second type of message associated with virtual goods A can be distinguished by message tag I, and another second type of message associated with virtual goods A can be distinguished by message tag II.

[0024] In one embodiment, after receiving multiple system messages, the message category and message tag of each system message are determined. By traversing the message category and message tag corresponding to each system message, it is determined whether the message category of the corresponding system message is the target category of the corresponding virtual product and whether the message tag is the target tag of the corresponding price. Then, system messages whose message category is the target category of the corresponding virtual product and whose message tag is the target tag of the corresponding price are identified as target messages. The target category is the message category of the corresponding virtual product, and the target tag is the message tag of the corresponding virtual product price. Optionally, system messages whose message category is the target category (such as the second type of message mentioned above) can be configured with corresponding message tags to correspond to different virtual products. Furthermore, the message tag can be used to determine whether the system message is associated with the price of the virtual product. For example, some system messages are associated with the listing or delisting of virtual product A, and some system messages are associated with changes in the price of virtual product A. These two types of system messages can be distinguished using different message tags. Therefore, by identifying the message category and message tag of system messages, this solution can quickly determine the target message from multiple received system messages, thereby improving the efficiency of predicting virtual product prices.

[0025] The process involves extracting text features from the identified target message to determine the corresponding text feature information. Optionally, the text feature information corresponds to key information contained in the text of the target message, such as associated virtual product names, price change keywords, etc., which are then used as text feature information. Optionally, a sliding window is a technique used in signal processing, image processing, and text processing. In the text processing process of extracting text features from the target message using a sliding window, the sliding window can extract n consecutive parameters (which can be characters, words, or other language units) from the text. It should be noted that in some embodiments, the text feature extraction of the target information can also be implemented using a neural network model built based on NLP (Natural Language Processing) algorithms. This model performs text preprocessing on the input information, such as word segmentation, stop word removal, and stemming, to convert the original text in the input information into a text form that can be further analyzed, and then extracts the corresponding text features from it.

[0026] Step S130: Based on the text feature information, determine the target virtual product and the price change trend of the target virtual product, and based on the price change trend, determine the predicted price of the target virtual product.

[0027] In one embodiment, after determining the text feature information, the target virtual product and its price change trend are determined based on the text feature information. The price change trend serves as a target parameter representing the price change information of the virtual product. Optionally, the price change trend can be determined by specific parameter values ​​of the target parameter, such as different parameter values ​​representing an increasing, decreasing, or stable trend. In one embodiment, key information is identified from the determined text feature information, and key information related to the virtual product is searched from the text feature information. For example, by comparing product keywords with the text feature information, the virtual product associated with the text feature information is determined, and this virtual product is then used as the target virtual product. Product keywords are keywords associated with the virtual product, such as the virtual product name and product identification number. Furthermore, the price change trend of the target virtual product is determined from the text feature information, such as by using a neural network model built based on NLP algorithms to determine keywords related to price changes in the text feature information, and then determining the price change trend through semantic analysis. Based on the price change trend, the price of the target virtual product in the next trading cycle is estimated and provided to the user as a predicted price to facilitate appropriate trading strategies. For example, a pre-set neural network model for predicting price changes outputs the corresponding predicted price. This neural network model is trained using price changes and actual transaction prices as sample data for each trading cycle, and it can predict prices that conform to the current price change trend.

[0028] As described above, this solution listens to the relevant message push channels and receives system messages, then extracts textual feature information from them. This textual feature information is used to determine the price change trend of the corresponding target virtual goods, thus combining system messages to judge virtual goods price changes. This makes the predicted prices provided to users more reliable. Furthermore, predicting virtual goods prices based on price change trends can combine the correlation between virtual goods and price changes, thereby improving the accuracy of virtual goods price predictions and facilitating the generation of accurate and reliable predicted prices as reference data.

[0029] In some embodiments, the message push channels include a first channel based on APNs (Apple Push Notification Service) and a second channel based on FCM (Firebase Cloud Messaging). Optionally, system messages are delivered based on message middleware in the server; that is, the server transmits system messages to different message push channels through message middleware to distribute system messages to user terminals with different operating system types. In this case, if the user terminal's operating system type is the first type, the user terminal determines the APNs-based first channel as the target channel among the preset message push channels, and receives the system messages transmitted by the server through the first channel. If the user terminal's operating system type is the second type, the user terminal determines the FCM-based second channel as the target channel among the preset message push channels. Therefore, through different message push channels, system messages can be delivered to the corresponding user terminals, enabling the user terminals to quickly receive the pushed system messages by listening to the corresponding target channel, thus facilitating real-time prediction of virtual product prices.

[0030] In one embodiment, system messages are transmitted via RocketMQ message middleware on the server. RocketMQ message middleware stores information about the corresponding message category through the `topic` subfield and information about the corresponding message tag through the `tag` subfield. When obtaining the message category and tag of a system message, for user terminals with an operating system type of 1, the received system message is pushed through a first channel. In this case, when system messages are pushed through the first channel, the user terminal extracts the target field corresponding to the user data in the system message by calling a preset APNs callback function, and obtains the message category from the `topic` subfield and the message tag from the `tag` subfield. Optionally, the APNs callback function processes messages received through the first channel by parsing the target field in the received message (such as the system message described above), which is used to store the `topic` and `tag` subfields. After the server establishes a connection with the user terminal, the server transmits system messages to the user terminal through the first channel, for example, in the form of a JSON string. This JSON string contains the message content of the pushed system message. When the server's message middleware transmits the system message through the first channel, it sets the message category and message tag in the `topic` and `tag` subfields, respectively, and stores them in the `target` field. The user terminal then calls the APNs callback function to retrieve the system message from the JSON string, and further retrieves the `topic` and `tag` subfields from the `target` field, determining the message category in the `topic` subfield and the message tag in the `tag` field.

[0031] For user terminals with an operating system type of 2, received system messages are pushed through a second channel. When system messages are pushed through the second channel, the user terminal parses the `data` field of the system message by calling a preset FCM callback function, and retrieves the message category from the `topic` subfield and the message tag from the `tag` subfield. Optionally, the FCM callback function processes messages received through the second channel; this function parses the `data` field of the system message, which stores the `topic` and `tag` subfields. Similarly, after the server establishes a connection with the user terminal, the server transmits system messages to the user terminal through the second channel, for example, in the form of a JSON string. This JSON string contains the message content of the pushed system message, and the server's message middleware sets the message category and message tag in the `topic` and `tag` subfields of the `data` field when transmitting system messages through the second channel. The user terminal then calls the FCM callback function to retrieve the system message from the JSON string, and further retrieves the `topic` and `tag` subfields from the `data` field, determining the message category in the `topic` subfield and the message tag in the `tag` field.

[0032] Therefore, according to the corresponding message push channel, the user terminal calls the corresponding callback function to extract the fields of the corresponding message category and message tag, so as to adapt to the system messages pushed by the message middleware, and quickly determine the message category and message tag of the system message, which helps to improve the recognition efficiency of system messages, and thus improve the prediction efficiency of price prediction.

[0033] Figure 2 The figure illustrates the steps for determining a target virtual product and its price change trend according to an embodiment of this application. In one embodiment, text feature information is determined by parsing the target message. The key information included in this text feature information can identify the target virtual product associated with the target message. Then, based on the product price knowledge graph, the price change trend of the corresponding target virtual product is determined from the text feature information. The specific steps are as follows: Step S210: Based on the preset product keywords of each virtual product, compare them one by one with the text feature information to obtain the relevance between each virtual product and the text feature information.

[0034] Step S220: Identify the virtual product with the highest relevance as the target virtual product.

[0035] Step S230: Determine the price change trend of the target virtual product based on the preset product price knowledge graph.

[0036] Optionally, product keywords are keywords associated with virtual products, such as the name of the virtual product, product identification number, etc. In one embodiment, multiple product keywords are configured for the same virtual product. These keywords are grouped according to different virtual products, and then each product keyword in each group is compared with text feature information to obtain the relevance between each virtual product and the text feature information. For example, in calculating the relevance, the number of times the product keyword corresponding to the same virtual product is hit in the text feature information can be determined during the comparison process. Then, the ratio of this number of hits to the total number of words in the text feature information is determined, and this ratio is used as the relevance. Furthermore, by comparing the relevance of each virtual product, the virtual product with the highest relevance is determined as the target virtual product.

[0037] A commodity price knowledge graph can record the relationships between different virtual commodities and price change trends. Furthermore, by analyzing these relationships, it can determine the price change trend of a virtual commodity caused by keywords related to price changes in system messages. For example, the commodity price knowledge graph records the relationship between a virtual commodity, keywords related to price changes, and the change trend using triples. Multiple triples are stored in the commodity price knowledge graph to represent various relationships. Based on this, by searching the commodity price knowledge graph according to textual feature information, the price change trend of a target virtual commodity can be determined.

[0038] In response, by determining the price change trend of the corresponding target virtual goods from text feature information, this solution can combine system messages to judge the price changes of virtual goods, thereby providing users with more reliable price predictions.

[0039] In one embodiment, the commodity price knowledge graph includes several triples associated with commodity entities, price keywords, and trends. These triples represent the relationships between commodity entities, price keywords, and trends. Optionally, a commodity entity corresponds to a virtual commodity, which is represented in the commodity price knowledge graph. The same commodity entity can be associated with several price keywords, and a price keyword is associated with a trend. A trend can be associated with at least one price keyword. Price keywords correspond to keywords that affect price changes, such as increased production or lower costs. Trends represent price changes, and different price keywords can correspond to the same trend. Figure 3 The figure illustrates the steps for determining a price change trend according to an embodiment of this application. By finding a target triplet, and then determining the price change trend of the target virtual commodity based on the changing trends within the target triplet, the specific steps are as follows: Step S310: Determine the target triplet containing the product entity corresponding to the target virtual product.

[0040] Step S320: Determine whether the text feature information of the corresponding target virtual product contains a target word that corresponds to the price keyword in the target triplet.

[0041] Step S330: If the target word is included, determine the price change trend of the target virtual product based on the change trend recorded in the target triplet.

[0042] Optionally, among the recorded triples, there may be multiple triples corresponding to the same product entity. In this case, all triples are initially screened according to the product entity corresponding to the target virtual product to determine the target triple. The product entity in this target triple is the product entity corresponding to the target virtual product. Then, the price keyword in the target triple is compared with the textual feature information of the corresponding target virtual product to determine whether the textual feature information contains a target word corresponding to the price keyword in the target triple. If the textual feature information is found to contain a target word, the trend associated with the price keyword is determined based on the correlation between the product entity, price keyword, and trend in the target triple. Based on the recorded trend, the price trend of the target virtual product is then determined.

[0043] For example, the determined textual feature information includes keywords such as "virtual product A" and "increased production." Accordingly, by searching for triples and filtering them, the triples corresponding to the virtual product A are selected as target triples. The textual feature information and the price keywords in the target triples are compared. For instance, if a target triple containing the corresponding virtual product A records a correlation between price keywords and price trends as "increased production - decreased price," then based on this correlation, it can be determined that an increase in the production of virtual product A will lead to a decrease in price. Therefore, the price trend of the target virtual product (i.e., virtual product A) can be determined to be a decreasing trend.

[0044] Therefore, this solution uses knowledge graphs to determine the relationship between virtual goods and price changes, thereby determining price change trends. This improves the accuracy of virtual goods price predictions and enhances the reliability of the predicted prices.

[0045] In one embodiment, the predicted price of a target virtual commodity can be output through a backpropagation (BP) neural network, thereby achieving fast and accurate price prediction and improving the reliability of the predicted price. Specifically, transaction price data from the previous two trading periods is obtained, such as by extracting transaction price data from historical data. This transaction price data and price change trends are then input into the configured BP neural network; that is, the transaction price data and price change trends are used as input data, and the predicted price of the target virtual commodity is output through the BP neural network.

[0046] Optionally, the BP neural network, as a multi-layer feedforward neural network, is trained using the backpropagation algorithm. The network structure of this BP neural network consists of an input layer, one or more hidden layers, and an output layer. It can handle non-linear relationships in the input data and perform prediction or classification. The input layer is the starting point of the BP neural network, responsible for receiving input data. Each neuron in the input layer corresponds to a feature value, and the input data enters the network through the neurons in the input layer. The hidden layers are responsible for extracting features from the input data and performing non-linear transformations. Each hidden layer consists of multiple neurons, and each neuron performs a weighted sum of the outputs of the previous layer and transforms the data using an activation function. For example, the output of the hidden layer is zj, and its calculation formula is:

[0047] Among them, w ji It is the weight connecting the i-th neuron in the input layer and the j-th neuron in the hidden layer, x i It is the output of the i-th neuron in the input layer, b j This is the bias term of the j-th neuron in the hidden layer. The corresponding activation function can be the Sigmoid function, whose formula is:

[0048] The output layer is used to generate the final prediction result, which can be represented by the following formula:

[0049] Among them, w kj It is the weight connecting the j-th neuron in the hidden layer and the k-th neuron in the output layer, a j It is the output of the hidden layer, b k It is the bias term of the k-th neuron in the output layer.

[0050] Furthermore, during model training, the model parameters of the BP neural network are optimized through a loss function, enabling the BP neural network to output predicted prices that better reflect price change trends. In this way, the BP neural network can determine the price change relationship caused by the price change trends of the previous two trading periods in the input data. Then, based on this non-linear relationship, it predicts the trading price brought about by the current price change trend, thereby determining the predicted price.

[0051] Figure 4 This is a schematic diagram of a virtual commodity price prediction device provided in an embodiment of this application. The device is used to execute the virtual commodity price prediction method provided in the above embodiment and has corresponding functional modules and beneficial effects for executing the method. As shown in the figure, the device includes a channel selection module 401, a message parsing module 402, and a price prediction module 403.

[0052] Among them, the channel selection module 401 is configured to determine the target channel as the monitoring target among the preset message push channels; The message parsing module 402 is configured to, when listening to multiple currently published system messages from the target channel, identify the target message associated with the virtual product among the multiple system messages, and parse the target message to obtain the corresponding text feature information; The price prediction module 403 is configured to determine the target virtual product and its price change trend based on text feature information, and to determine the predicted price of the target virtual product based on the price change trend.

[0053] Based on the above embodiments, the message parsing module 402 is specifically configured as follows: Determine the message category and message tag for each system message; System messages whose message category is the target category of the corresponding virtual product and whose message tag is the target tag of the corresponding price are identified as target messages.

[0054] Based on the above embodiments, the message push channel includes a first channel based on APNs and a second channel based on FCM. System messages are transmitted based on the RocketMQ message middleware in the server. The RocketMQ message middleware stores information about the corresponding message category through the topic subfield and information about the corresponding message tag through the tag subfield. The message parsing module 402 is also configured as follows: When system messages are pushed through the first channel, the target fields of the corresponding user data in the system message are extracted by calling the preset APNs callback function, and the message category in the topic subfield and the message tag in the tag subfield are obtained from it. The APNs callback function is used to process messages received through the first channel. When system messages are pushed through the second channel, the data field in the system message is parsed by calling the preset FCM callback function, and the message category in the topic subfield and the message tag in the tag subfield are obtained from it. The FCM callback function is used to process messages received through the second channel.

[0055] Based on the above embodiments, the price prediction module 403 is specifically configured as follows: Based on the preset product keywords of each virtual product, the relevance of each virtual product to the text feature information is compared one by one. The virtual goods with the highest relevance are identified as the target virtual goods; Based on a pre-defined knowledge graph of commodity prices, determine the price change trend of the target virtual commodity.

[0056] Based on the above embodiments, the commodity price knowledge graph includes several triples associated with commodity entities, price keywords, and change trends. The price prediction module 403 is also configured as follows: Determine the target triple containing the product entity corresponding to the target virtual product; Determine whether the text feature information of the corresponding target virtual product contains target words that correspond to the price keywords in the target triple; When the target term is included, the price change trend of the target virtual product is determined based on the change trend recorded in the target triple.

[0057] Based on the above embodiments, the channel selection module 401 is specifically configured as follows: When the user terminal's operating system type is type 1, the first channel based on APNs is determined as the target channel in the preset message push channels; When the user terminal's operating system type is type 2, the second channel based on FCM is selected as the target channel in the preset message push channels.

[0058] Based on the above embodiments, the grid prediction module 403 is further configured as follows: Obtain the trading price data from the previous two trading periods, and input the trading price data and price change trend into the set BP neural network; The predicted price of the target virtual product is output through a backpropagation neural network.

[0059] It is worth noting that in the embodiments of the above-mentioned device, the modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of each module are only for easy differentiation and are not used to limit the protection scope of the embodiments of this application.

[0060] Figure 5 This is a schematic diagram of an electronic device provided in an embodiment of this application. The device is used to execute the virtual commodity price prediction method provided in the above embodiment and has corresponding functional modules and beneficial effects for executing the method. As shown in the figure, the device includes a processor 501, a memory 502, an input device 503, and an output device 504. The number of processors 501 can be one or more; one processor 501 is shown as an example in the figure. The processor 501, memory 502, input device 503, and output device 504 can be connected via a bus or other means; a bus connection is shown as an example in the figure. The memory 502, as a computer-readable storage medium, can be used to store software programs, computer-executable programs, and modules, such as the program instructions / modules corresponding to the virtual commodity price prediction method in the embodiments of this application. The processor 501 executes various corresponding functional applications and data processing by running the software programs, instructions, and modules stored in the memory 502, thereby realizing the above-mentioned virtual commodity price prediction method.

[0061] The memory 502 may primarily include a program storage area and a data storage area. The program storage area may store the operating system and applications required for at least one function; the data storage area may store data recorded or created during use. Furthermore, the memory 502 may include high-speed random access memory and non-volatile memory, such as at least one disk storage device, flash memory device, or other non-volatile solid-state storage device. In some embodiments, the memory 502 may further include memory remotely configured relative to the processor 501, which can be connected to the device via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0062] The input device 503 can be used to input corresponding digital or character information to the processor 501, and to generate key signal inputs related to the user settings and function control of the device; the output device 504 can be used to send or display key signal outputs related to the user settings and function control of the device.

[0063] This application also provides a storage medium storing computer-executable instructions, which, when executed by a processor, are used to perform relevant operations in the virtual commodity price prediction method provided in any embodiment of this application.

[0064] Computer-readable storage media include both permanent and non-permanent, removable and non-removable media, and information storage can be achieved by any method or technology. Information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, disk storage or other magnetic storage devices, or any other non-transfer medium that can be used to store information accessible by a computing device.

[0065] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0066] Note that the above description is merely a preferred embodiment and the technical principles employed in this application. Those skilled in the art will understand that this application is not limited to the specific embodiments described herein, and various obvious changes, readjustments, and substitutions can be made without departing from the scope of protection of this application. Therefore, although this application has been described in detail through the above embodiments, this application is not limited to the above embodiments. Many other equivalent embodiments may be included without departing from the concept of this application, and the scope of this application is determined by the scope of the appended claims.

Claims

1. A method for predicting the price of virtual goods, characterized in that, include: Select the target channel as the monitoring target from the preset message push channels; When multiple system messages are detected from the target channel, a target message associated with virtual goods is identified from the multiple system messages, and the target message is parsed to obtain the corresponding text feature information. Based on the text feature information, the target virtual product and its price change trend are determined, and based on the price change trend, the predicted price of the target virtual product is determined.

2. The virtual goods price prediction method according to claim 1, characterized in that, The step of determining the target message associated with the virtual product among the multiple system messages includes: Determine the message category and message tag for each of the system messages; System messages whose message category is the target category of the corresponding virtual product and whose message tag is the target tag of the corresponding price are identified as target messages.

3. The virtual goods price prediction method according to claim 2, characterized in that, The message push channel includes a first channel based on APNs and a second channel based on FCM. The system messages are transmitted based on the RocketMQ message middleware in the server. The RocketMQ message middleware stores information of the corresponding message category through the topic subfield and information of the corresponding message tag through the tag subfield. Determining the message category and message tag for each system message includes: When the system message is pushed through the first channel, the target field of the corresponding user data in the system message is extracted by calling the preset APNs callback function, and the message category in the topic subfield and the message tag in the tag subfield are obtained from it. The APNs callback function is used to process the message received through the first channel. When the system message is pushed through the second channel, the data field in the system message is parsed by calling a preset FCM callback function, and the message category in the topic subfield and the message tag in the tag subfield are obtained from it. The FCM callback function is used to process the message received through the second channel.

4. The virtual goods price prediction method according to claim 1, characterized in that, The step of determining the target virtual product and its price change trend based on the text feature information includes: Based on the preset product keywords of each virtual product, the relevance of each virtual product to the text feature information is obtained by comparing them one by one. The virtual goods with the highest relevance are identified as the target virtual goods; Based on a pre-defined commodity price knowledge graph, the price change trend of the target virtual commodity is determined.

5. The virtual goods price prediction method according to claim 4, characterized in that, The commodity price knowledge graph includes several triples that are associated with commodity entities, price keywords, and price trends; The step of determining the price change trend of the target virtual product based on a preset product price knowledge graph includes: Determine the target triple containing the product entity corresponding to the target virtual product; Determine whether the text feature information corresponding to the target virtual product contains a target word that corresponds to the price keyword in the target triple; In the case of the target word, the price change trend of the target virtual product is determined based on the change trend recorded in the target triple.

6. The virtual commodity price prediction method according to any one of claims 1-5, characterized in that, The step of determining the target channel as the monitoring target in the preset message push channels includes: When the operating system type of the user terminal is the first type, the first channel based on APNs is determined as the target channel in the preset message push channels; When the operating system type of the user terminal is the second type, the second channel based on FCM is determined as the target channel in the preset message push channel.

7. The virtual commodity price prediction method according to any one of claims 1-5, characterized in that, Determining the predicted price of the target virtual commodity based on the price change trend includes: Obtain the transaction price data from the previous two trading periods, and input the transaction price data and the price change trend into the set BP neural network; The predicted price of the target virtual product is output through the BP neural network.

8. A virtual commodity price prediction device, characterized in that, include: The channel selection module is configured to determine the target channel as the monitoring target among preset message push channels; The message parsing module is configured to, when multiple system messages are currently being monitored from the target channel, determine the target message associated with the virtual product among the multiple system messages, and parse the target message to obtain the corresponding text feature information; The price prediction module is configured to determine the target virtual product and its price change trend based on the text feature information, and to determine the predicted price of the target virtual product based on the price change trend.

9. An electronic device, characterized in that, include: One or more processors; A storage device for storing one or more programs that, when executed by one or more processors, cause the one or more processors to implement the virtual commodity price prediction method as described in any one of claims 1-7.

10. A storage medium for storing computer-executable instructions, characterized in that, The computer-executable instructions, when executed by a processor, are used to perform the virtual commodity price prediction method as described in any one of claims 1-7.

Images