A method and apparatus for generating a prediction model of supply and demand changes for game items.

Abstract

This application discloses a method for generating a prediction model of game item supply and demand changes, comprising: obtaining historical supply and demand change data of game items in a game scene; determining the time category corresponding to the generation date of the historical supply and demand change data; obtaining first change data within a first preset time period and second change data within a second preset time period from the historical supply and demand change data, wherein the first preset time period is earlier than the second preset time period; using the first change data and the second change data as training samples to train a preset initial prediction model of game item supply and demand changes, thereby generating a prediction model corresponding to the time category for predicting the supply and demand changes of game items in the game scene.

Description

Technical Field

[0001] This application relates to the field of data processing, specifically to a method and apparatus for generating a prediction model of supply and demand changes for game items. This application also relates to a method and apparatus for determining abnormal situations of supply and demand changes for game items. This application also relates to an electronic device and a computer storage medium. Background Technology

[0002] With the continuous development of computer networks, more and more people are choosing to relieve life's pressures by playing online games. In games, players can obtain in-game items through gameplay and exchange these items with other players via in-game trading platforms.

[0003] Based on this game mode, a large number of accounts are created for profit. These game characters are usually operated by illegal scripts that are not manually or semi-manually controlled. They continuously acquire a large number of game items through game behavior, disrupting the originally fair item exchange rules in the game scene and seriously affecting the game experience of other players.

[0004] Therefore, how to create a fair environment for exchanging items in games and improve the player's gaming experience has become a technical problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0005] This application provides a method, apparatus, electronic device, and computer storage medium for generating a prediction model of game item supply and demand changes, to solve the technical problems existing in the prior art. This application also provides a method, apparatus, electronic device, and computer storage medium for determining abnormal situations of game item supply and demand changes.

[0006] The method for generating a prediction model for the supply and demand changes of game items provided in this application includes:

[0007] Obtain historical supply and demand change data for at least one game item in the game scene, wherein the historical supply and demand change data corresponds to the generation date information of the historical supply and demand change data;

[0008] Determine the time category corresponding to the generation date of the historical supply and demand change data;

[0009] From the historical supply and demand change data, obtain first change data within a first preset time period and second change data within a second preset time period, wherein the first preset time period is earlier than the second preset time period;

[0010] The first change data and the second change data are used as training samples to train a preset initial game item supply and demand change prediction model, thereby generating a prediction model corresponding to the time category for predicting game item supply and demand changes within the game scene.

[0011] Optionally, obtaining historical supply and demand data for game items in the game scene includes:

[0012] Obtain data on the changes in the total amount of game items and the transaction data of game items generated by the historical game behavior of game players in the first game server, wherein the generation date of the historical game behavior of the game players is the same as the generation date of the historical supply and demand change data;

[0013] Based on the changes in the total quantity of game items and the transaction data of game items, the historical supply and demand changes of game items within the game scene are determined.

[0014] Optionally, obtaining historical supply and demand data for game items within the game scene includes:

[0015] Obtain data on the changes in the total amount of tradable game items listed on the second game server and data on game item transactions generated by players' game behavior, wherein the generation date of the players' game behavior is the same as the generation date of the historical supply and demand change data;

[0016] Based on the quantity of the tradable game items and the game item trading data, determine the historical supply and demand changes of game items within the game scene.

[0017] Optionally, determining the time category corresponding to the generation date of the historical supply and demand change data includes:

[0018] Obtain other supply and demand change data corresponding to different time categories;

[0019] Obtain the first data change characteristics of other supply and demand change data corresponding to each different time category and the second data change characteristics of the historical supply and demand change data;

[0020] Based on the first similarity between the first data change feature and the second data change feature, the time category corresponding to the generation date of the historical supply and demand change data is determined.

[0021] Optionally, the method further includes:

[0022] Preset date and corresponding time category;

[0023] The time category includes at least one of the following: regular weekday, regular holiday, game update day.

[0024] Optionally, obtaining the first data change characteristics of other supply and demand change data corresponding to each different time category and the second data change characteristics of the historical supply and demand change data includes:

[0025] The historical supply and demand change data is divided into time nodes according to a preset first time interval, and the supply and demand change characteristics at each time node of the historical supply and demand change data are determined as the first data change characteristics.

[0026] The other supply and demand change data are divided into time nodes according to the first time interval, and the supply and demand change characteristics at each node of the other supply and demand change data are determined as the second data change characteristics.

[0027] Optionally, determining the time category corresponding to the generation date of the historical supply and demand change data based on the similarity between the first data change feature and the second data change feature includes:

[0028] Based on the second similarity between the first data change characteristics and the second data change characteristics at each same time point, the first similarity between the historical supply and demand change data and other supply and demand change data corresponding to each time category is determined respectively;

[0029] The time category corresponding to the other supply and demand change data whose first similarity is greater than the preset first similarity threshold is taken as the time category corresponding to the generation date of the historical supply and demand change data.

[0030] Optionally, obtaining the first change data within a first preset time period and the second change data within a second preset time period from the historical supply and demand change data includes:

[0031] Obtain the first preset time period and the second preset time period set for the initial game item supply and demand change prediction model;

[0032] Extract the first change data corresponding to the first preset time period from the historical supply and demand change data;

[0033] Extract the second change data corresponding to the second preset time period from the historical supply and demand change data.

[0034] Optionally, the step of using the first change data and the second change data as training samples to train a preset initial game item supply and demand change prediction model, generating a prediction model corresponding to the time category for predicting game item supply and demand changes within the game scene, includes:

[0035] The first change data is input into the initial game item supply and demand change prediction model to obtain the first prediction data of the game item supply and demand change data in the second time period output by the initial game item supply and demand change prediction model.

[0036] If the second similarity between the second change data and the predicted data is greater than a preset second similarity threshold, the initial game item supply and demand change prediction model is trained based on the first change data and the second change data to obtain the game item supply and demand change prediction model.

[0037] Optionally, the method further includes:

[0038] If the similarity between the second changed data and the first predicted data is less than or equal to a preset second similarity threshold, the duration of the first preset time period and the second preset time period is adjusted according to the adjustment range of the first preset time period and the second preset time period.

[0039] The supply and demand change data corresponding to the first preset time period after adjusting the time length is used as the first change data, and the supply and demand change data corresponding to the second preset time period after adjusting the time length is used as the second change data. The process then returns to the step of inputting the first change data into the initial game item supply and demand change prediction model to obtain the predicted data for the game item supply and demand change data within the second time period output by the initial game item supply and demand change prediction model. This application also provides a method for determining abnormal situations in game item supply and demand changes, including:

[0040] Obtain first supply and demand change data of game items within a first preset time period and second supply and demand change data of game items within a second preset time period, wherein the time category corresponding to the generation date of the first supply and demand change data is the same as the time category corresponding to the generation date of the second supply and demand change data.

[0041] The first supply and demand change data is input into the game item supply and demand change prediction model corresponding to the time category to obtain the predicted data of game item supply and demand change data within the preset second time period output by the game item supply and demand change prediction model.

[0042] Based on the similarity between the second supply and demand change data and the predicted data, abnormal situations in the supply and demand changes of game items within the second time period are determined.

[0043] Optionally, determining abnormal situations in the supply and demand of game items within the second time period based on the similarity between the second supply and demand change data and the predicted data includes:

[0044] Determine whether the similarity between the second supply and demand change data and the preset data is less than a preset similarity threshold;

[0045] If the similarity between the second supply and demand change data and the preset data is less than the preset similarity threshold, then it is determined that the supply and demand change of game items in the second time period is abnormal.

[0046] If the similarity between the second supply and demand change data and the preset data is greater than or equal to the preset similarity threshold, then it is determined that the supply and demand change of game items within the second time period is normal.

[0047] Optionally, the method further includes:

[0048] Preset date and corresponding time category;

[0049] The time category includes at least one of the following: regular weekday, regular holiday, game update day.

[0050] This application also provides a device for generating a prediction model for the supply and demand changes of game items, including:

[0051] The first acquisition unit is used to acquire historical supply and demand change data of game items in the game scene;

[0052] A determining unit is used to determine the time category corresponding to the generation date of the historical supply and demand change data;

[0053] The second obtaining unit is used to obtain first change data within a first preset time period and second change data within a second preset time period from the historical supply and demand change data, wherein the first preset time period is earlier than the second preset time period.

[0054] The generation unit is used to train a preset initial game item supply and demand change prediction model using the first change data and the second change data as training samples, and to generate a prediction model corresponding to the time category for predicting game item supply and demand changes in the game scene.

[0055] This application also provides a device for determining abnormal changes in the supply and demand of game items, comprising:

[0056] The third obtaining unit is used to obtain first supply and demand change data of game items within a first preset time period and second supply and demand change data of game items within a second preset time period, wherein the time category corresponding to the generation date of the first supply and demand change data is the same as the time category corresponding to the generation date of the second supply and demand change data.

[0057] The model prediction unit is used to input the first supply and demand change data into the game item supply and demand change prediction model corresponding to the time category, and obtain the prediction data of the game item supply and demand change data within a preset second time period output by the game item supply and demand change prediction model.

[0058] An anomaly detection unit is used to determine anomalies in the supply and demand of game items within a second time period based on the similarity between the second supply and demand change data and the predicted data.

[0059] This application also provides an electronic device, including:

[0060] processor;

[0061] A memory for storing a program of a method, which, when read and executed by a processor, performs any of the methods described above.

[0062] This application also provides a computer storage medium storing a computer program, which, when executed, performs any of the methods described above.

[0063] Compared with the prior art, this application has the following advantages:

[0064] The method for generating a game item supply and demand change prediction model provided in this application includes: obtaining historical supply and demand change data of game items in a game scene; determining the time category corresponding to the generation date of the historical supply and demand change data; obtaining first change data within a first preset time period and second change data within a second preset time period from the historical supply and demand change data, wherein the first preset time period is earlier than the second preset time period; using the first change data and the second change data as training samples to train a preset initial game item supply and demand change prediction model, thereby generating a prediction model corresponding to the time category for predicting game item supply and demand changes in the game scene.

[0065] This method categorizes the historical supply and demand change data by generation date and constructs a predictive model corresponding to the time category, based on the change data at different points in time within the historical supply and demand change data, to predict the supply and demand changes of game items within the game scene. This enables the prediction of normal supply and demand change data within the game scene. The model obtained by this method can accurately predict the normal supply and demand relationship of game items within the game scene. The predicted supply and demand change data within a preset time period obtained by the model provides a reference for the actual supply and demand change data generated within the game scene, enabling game maintenance personnel to monitor abnormal time nodes within the game scene and then perform effective management operations for the abnormal time nodes. Attached Figure Description

[0066] Figure 1a A schematic diagram illustrating an application scenario of a game item supply and demand change prediction model, as provided in one embodiment of this application;

[0067] Figure 1 A flowchart illustrating a method for generating a game item supply and demand change prediction model, as provided in another embodiment of this application.

[0068] Figure 2 A timeline diagram illustrating the supply and demand changes of game items provided in another embodiment of this application;

[0069] Figure 3 A schematic diagram of a supply and demand change data center point provided in another embodiment of this application;

[0070] Figure 4 A flowchart illustrating the method for determining abnormal supply and demand changes of game items provided in this application embodiment;

[0071] Figure 5 A schematic diagram of the structure of a device for generating a game item supply and demand change prediction model provided in another embodiment of this application;

[0072] Figure 6 A schematic diagram of a device for determining abnormal supply and demand of game items, provided in another embodiment of this application;

[0073] Figure 7 This is a schematic diagram of an electronic device structure provided for another embodiment of this application. Detailed Implementation

[0074] Numerous specific details are set forth in the following description to provide a full understanding of the invention. However, this application can be implemented in many other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of this application; therefore, this application is not limited to the specific embodiments disclosed below.

[0075] To facilitate understanding of the method for generating the game item supply and demand change prediction model provided in this application, this application first provides a specific application scenario of the model, and then introduces the application of the game item supply and demand change prediction model in conjunction with this scenario.

[0076] Please refer to Figure 1a , Figure 1a This is a schematic diagram illustrating an application scenario of a game item supply and demand change prediction model, provided as an embodiment of this application.

[0077] like Figure 1a As shown, Figure 1 It includes: game database 101, data classification unit 102, model processing unit 103, and anomaly determination unit 104.

[0078] The game database 101 can be understood as a database used to store game data for a specific game. In the scenario embodiment of this application, the game database is mainly used to store supply and demand change data of game items stored on a daily basis. For example, for game A, the game database 101 stores the supply and demand change data of game items in the first game server area of ​​game A over the past 10 days, as well as the current real-time supply and demand change data of game items in the first game server area.

[0079] The supply and demand data for game items can be understood as the relationship between the number of game items that players list for trading in the game auction house and the number of game items that players purchase through the auction house (wherein, the act of players listing game items can be regarded as an increase in the supply of game items, and the act of players removing game items from the auction house or purchasing game items can be regarded as a decrease in the supply of game items).

[0080] The data classification unit 102 is used to obtain the supply and demand change data of game items from the game database 101 on a daily basis, and classify the supply and demand change data of game items according to the date.

[0081] In the scenario embodiments provided in this application, the categorized types may include: game version update date, regular weekday, and regular weekend. That is, the time category of the game item supply and demand change data generation time is divided according to the generation date of the game item supply and demand change data.

[0082] For example, suppose that game A updates every Thursday, and the data classification unit 102 obtains game item supply and demand change data from January 1, 2022 to January 7, 2022 from the game database 101 (where January 1 and 2, 2022 are weekends, January 6, 2022 is a Thursday, and the rest of the time are normal working days), which are: Data 1, Data 2, Data 3, Data 4, Data 5, Data 6, and Data 7. The specific classification of the game item supply and demand change data obtained within the above dates is shown in the following table:

[0083] Game version update date ordinary working day ordinary weekend Data 6 Data 3, Data 4, Data 5 Data 1, Data 2

[0084] In one optional embodiment of this application, the above-described classification process of the game item supply and demand change data can be carried out manually or in other ways. Since this application only describes the application scenario and process of the game item supply and demand change prediction model, the detailed process of classifying the game item supply and demand change data can be found in the relevant descriptions of other embodiments of this application, and will not be repeated here.

[0085] The model processing unit 103 is used to store trained prediction models for game item supply and demand changes, each corresponding to a different time category. These prediction models are used to predict game item supply and demand changes in a second time period based on game item supply and demand change data within a first time period in the game scene. The game item supply and demand change data for both the first and second time periods are generated on the same date, but the generation time of the game item supply and demand change data for the first time period is earlier than the generation time of the game item supply and demand change data for the second time period.

[0086] In one optional application scenario, assuming the game maintenance personnel want to predict abnormal time periods of supply and demand changes for game items on January 1, 2022, they can extract the first game item supply and demand change data (hereinafter referred to as the first data) within the first time period on January 1, 2022, and the second game item supply and demand change data (hereinafter referred to as the second data) within the second time period. The first data is then sent to the model processing unit 103, and the second data is sent to the anomaly determination unit 104.

[0087] The model processing unit 103 uses the prediction model to predict the supply and demand changes of game items in the second time period, obtains the predicted data for the supply and demand changes of game items in the second time period, and sends the predicted data to the anomaly determination unit 104.

[0088] The anomaly determination unit 104 is used to determine the correlation between the abnormal data and the predicted data. If the correlation between the two is greater than or equal to a predetermined correlation threshold, it is determined that the supply and demand relationship of game items in the second time period is not abnormal; if the correlation between the two is less than a predetermined correlation threshold, it is determined that the second time period is an abnormal time period. At this time, relevant game maintenance personnel need to investigate the players in the second time period to determine the source of the anomaly and address it. For example, assuming the anomaly is caused by a studio using scripts to maliciously disrupt the fair play environment with a large number of accounts, the accounts involved in the studio need to be banned.

[0089] It should be noted that the above description is only for the purpose of understanding one possible application scenario of the game item supply and demand change prediction model provided in this application, and is not intended to limit the scenario. The game item supply and demand change prediction model provided in this application can also be used in other game scenarios.

[0090] Another embodiment of this application provides a method for generating a prediction model of supply and demand changes for game items. Please refer to [link / reference]. Figure 1 , Figure 1 A flowchart illustrating a method for generating a prediction model for the supply and demand of game items, as provided in another embodiment of this application.

[0091] like Figure 1 As shown, the method for generating the game item supply and demand change prediction model includes the following steps S101 to S104.

[0092] Step S101: Obtain historical supply and demand data for game items in the game scene;

[0093] In this embodiment, the game items in the game scene can be understood as tradable items within the game scene. For example, for MMORPGs (Massive Multiplayer Online Role-Playing Games), the game items in the game scene can be game items obtained by players when performing in-game tasks or other game activities, and can be traded in the in-game auction house or on an online item trading platform specifically designed for the game.

[0094] The historical supply and demand data of the game items can be understood as the relationship between the production and purchase of the game items in the game scene, or the relationship between the number of game items listed and sold on the trading platform (which can be a trading platform within the game scene or other online trading platforms outside the game scene).

[0095] In one optional embodiment of this application, step S101 can be implemented by steps S1 and S2:

[0096] Step S1: Obtain data on the change in the total amount of game items generated by the historical game behavior of game players in the first game server and data on game item transactions, wherein the generation date of the historical game behavior of the game players is the same as the generation date of the historical supply and demand change data;

[0097] Step S2: Based on the change data of the total amount of game items and the transaction data of game items, determine the historical supply and demand change data of game items in the game scene.

[0098] In this embodiment, the first game server refers to a server in a certain region of the game. The first game server can be selected according to the actual situation. In one optional embodiment of this application, the first game server refers to a server in which the supply and demand of game items tends to be stable. Generally, a game region server with a long opening time or a special server in the game (for example, a test server specifically used for testing new game versions) can be selected.

[0099] The historical supply and demand change data refers to the supply and demand change data generated based on the historical game behavior of game players. For example, if game players obtain game items through game behavior, it means that the supply and demand change data of the game items is increasing. If game players purchase game items or make game items untradeable through game behavior, it means that the supply and demand change data of the game items is decreasing.

[0100] It should be noted that the historical supply and demand data of game items in this embodiment are obtained from the game database, which contains several historical supply and demand data points on a daily basis.

[0101] In another optional embodiment of this application, step S101 can also be implemented by steps S3 and S4.

[0102] Step S3: Obtain data on the change in the total amount of tradable game items listed in the second game server and data on game item transactions generated by the game players' game behavior, wherein the generation date of the game players' game behavior is the same as the generation date of the historical supply and demand change data;

[0103] Similar to the first game server mentioned above, the second game server is also used to refer to a server in a certain region of the game. For details, please refer to the above introduction of the first game server. It will not be repeated here.

[0104] In step S3, the total amount of tradable game items refers to the sum of game items listed by game players through the in-game auction house or store open to game players; the game item transaction data refers to the number of game items purchased by game players through the auction house or store.

[0105] Step S4: Based on the quantity of the tradable game items and the game item trading data, determine the historical supply and demand change data of the game items in the game scene.

[0106] In this embodiment, the act of a player listing game items represents an increase in the supply and demand of those game items, while the act of a player purchasing game items represents a decrease in the supply and demand of those game items. In other words, the historical supply and demand data mentioned in step S4 above is the supply and demand curve plotted based on the aforementioned data of increasing and decreasing supply and demand.

[0107] Step S102: Determine the time category corresponding to the generation date of the historical supply and demand change data;

[0108] In practical applications, player demand for games varies at different times (on a daily basis). Therefore, the supply and demand data for game items in different game environments also exhibit different characteristics at different times. For example, on normal working days, the number of players in the game is relatively small, naturally resulting in lower production and trading volumes of game items. Conversely, on weekends, the number of players in the game is relatively large, leading to higher production and trading volumes of game items. Therefore, to accurately grasp the characteristics of game item supply and demand at different times, it is first necessary to categorize the historical supply and demand change data obtained in step S101 by date.

[0109] In one optional embodiment of this application, the time category includes, but is not limited to, at least one of the following categories: ordinary weekdays, ordinary holidays, game update days, and Spring Festival.

[0110] In practical applications, the classification of the historical supply and demand change data generation dates can be done manually. For example, a table of historical supply and demand change data generation dates can be drawn by referring to a calendar to determine the time category corresponding to the generation date.

[0111] In another optional embodiment of this application, it is assumed that the generation time of the supply and demand change data of game items has been classified, and other supply and demand change data exist in each time category. Then, the time category of the historical supply and demand change data can be determined based on the data characteristics of the other supply and demand change data. Specifically, step S102 can also be implemented through the following steps S5 to S7:

[0112] Step S5: Obtain other supply and demand change data corresponding to each different time category;

[0113] Step S6: Obtain the first data change characteristics of other supply and demand change data corresponding to each different time category and the second data change characteristics of the historical supply and demand change data;

[0114] In one optional embodiment of this application, the first change feature data is composed of node features of the other supply and demand change data in different preset time periods, and the corresponding second data change feature is composed of node features of the historical supply and demand change data in the preset time period.

[0115] Please refer to Figure 2 , Figure 2 A timeline diagram illustrating the supply and demand changes of game items provided in another embodiment of this application.

[0116] like Figure 2Assuming that the extracted historical supply and demand change data of game items (supply and demand change data 0) is the supply and demand change data in the game scene from 12:00 to 16:00 on a certain date, and other historical supply and demand change data (supply and demand change data 1, supply and demand change data 2) is other supply and demand change data in the game scene from 12:01 to 18:00 on other dates, then the historical supply and demand change data and other supply and demand change data can be divided into nodes with 1 hour as a time node.

[0117] That is, for the historical supply and demand change data, the supply and demand change data corresponding to the time period 12:01-13:00 is the supply and demand change data of the first node; the supply and demand change data corresponding to the time period 13:01-14:00 is the supply and demand change data of the second node; the supply and demand change data corresponding to the time period 14:01-15:00 is the supply and demand change data of the third node; and the supply and demand change data corresponding to the time period 15:01-16:00 is the supply and demand change data of the fourth node.

[0118] Then, extract the data change characteristics a1 of the supply and demand change data of the first node, a2 of the supply and demand change data of the second node, a3 of the supply and demand change data of the third node, and a4 of the supply and demand change data of the fourth node.

[0119] Furthermore, the same processing is performed on the other supply and demand change data to divide each other supply and demand change data into nodes, and to determine the data change characteristics b1, b2, b3, and b4 at the other supply and demand change data nodes.

[0120] Step S7: Determine the time category corresponding to the generation date of the historical supply and demand change data based on the first similarity between the first data change feature and the second data change feature.

[0121] Specifically, step S7 above includes:

[0122] Based on the similarity between the first change data feature and the second data change feature at each same time point, the first similarity between the historical supply and demand change data and other supply and demand change data corresponding to each time category is determined respectively;

[0123] Specifically, the first similarity can be represented by the distance between the first change data feature and the second data change feature at each same time point. For example, suppose the supply and demand change features at each time point of the historical supply and demand change data are: a1, a2, a3, ... a m The supply and demand change characteristics at each time point of the other supply and demand change data are b1, b2, b3, ... b m .

[0124] The first similarity can then be obtained using the following formula (1):

[0125]

[0126] Here, Similarity represents the first similarity.

[0127] Finally, the time category corresponding to the other supply and demand change data whose first similarity is greater than a preset similarity threshold is taken as the time category corresponding to the generation date of the historical supply and demand change data.

[0128] In another optional embodiment of this application, if the other supply and demand change data are not classified in time, after finding other supply and demand change data similar to the historical supply and demand change data, it is necessary to further classify the other supply and demand change data with the historical supply and demand change data through clustering.

[0129] Specifically, after obtaining the similarity between the historical supply and demand change data and the other supply and demand change data (i.e., after executing Formula 1 above), it is necessary to further calculate the pairwise similarity between each of the other supply and demand change data. Specifically, the calculation of the pairwise similarity between each of the other supply and demand change data can still be achieved using the above Formula (1).

[0130] Furthermore, in this embodiment, the KMeans clustering algorithm can be used to classify the historical supply and demand change data and other supply and demand change data. The process of classifying the change data using the above algorithm includes the following steps S8 to S10:

[0131] To facilitate understanding of the iterative calculation process of the KMeans clustering algorithm provided in this application, firstly, the similarity between different supply and demand change data can be understood as the distance between different supply and demand change data. Specifically, please refer to the above formula (1). In practical applications, the supply and demand change characteristics at each time point can be represented in the form of a vector. For example, for the historical supply and demand change data, the supply and demand change characteristic a at a certain time point is... n (n = 1, 2, 3…n) can be understood as a vector, for example, a1 = [1, 2, 3]. Therefore, the method of calculating similarity provided by formula (1) in this application can actually be understood as the distance between different vectors.

[0132] That is, the above formula (1) can also be expressed in the following way:

[0133]

[0134] Step S8: Randomly select benchmark supply and demand change data from historical supply and demand change data obtained by date and other supply and demand change data.

[0135] Step S9: Calculate the distance between the baseline supply and demand change data and the historical supply and demand change data and the other supply and demand change data, and determine the same number of distance center points as the number of time categories;

[0136] Step S10: Supply and demand change data within a preset distance range from the center point are classified as supply and demand change data of the same time category.

[0137] Please refer to Figure 3 , Figure 3 A schematic diagram of a supply and demand change data center provided for another embodiment of this application.

[0138] Figure 3 It includes data 1 to data 10, as well as randomly selected data c (benchmark supply and demand change data).

[0139] Assuming that data points 0-10 away from data point c belong to one time category, and data points 30-50 away from data point c belong to another time category, then data points 1 to 10 can be divided into two categories based on the distance between them. One category is centered on data point c (Category 1), and the data points are 0-10 away from data point c (Category 2). Figure 3 The first category consists of data 1 to data 5, while the second category has a distance of 30 to 50 from data c, and the pairwise distance between any two data points in the second category is also less than 10 (in the context of...). Figure 3 The data in the middle ranges from 6 to 10, and the center point is the distance between data 8 and data c, which is 40.

[0140] In another embodiment of this application, in order to simplify the implementation of the algorithm, step 9 above can also be modified into step S9-1: calculate the distance between the benchmark supply and demand change data and the historical supply and demand change data and the other supply and demand change data, and determine the center point corresponding to the two time categories (first time category and second time category).

[0141] Then, in step 10, the supply and demand change data in the first time category and the second time category are determined respectively.

[0142] For other time categories, select the remaining historical supply and demand change data and other supply and demand change data. Return to steps S9-1 and 10 until the preset classification criteria are met.

[0143] In this embodiment of the application, the preset classification criterion may be that the ratio of different categories of supply and demand change data determined by step S10 is less than a preset ratio threshold; or, the number of different categories of supply and demand change data determined by step S10 is less than a preset first quantity threshold; or, the sum of the remaining historical supply and demand change data and other supply and demand change data is less than a preset second quantity threshold.

[0144] Step S103: Obtain first change data within a first preset time period and second change data within a second preset time period from the historical supply and demand change data, wherein the first preset time period is earlier than the second preset time period;

[0145] Obtaining the first change data within a first preset time period and the second change data within a second preset time period from the historical supply and demand change data involves extracting node features at preset nodes in the historical supply and demand change data. For example, assuming the historical supply and demand change data includes the supply and demand change data of game items in a game scene from 12:01 to 18:00 on a certain date, and the historical supply and demand change data is divided into 1-hour time nodes, then the first preset time period can be 12:00 to 14:00, the second time period can be 14:01 to 15:00, the first change data can be the supply and demand change data corresponding to the time range of 12:00 to 14:00, and the second change data can be the supply and demand change data corresponding to the time range of 14:01 to 15:00.

[0146] To facilitate understanding of the application of the first change data in the first preset time period and the second change data in the second preset time period during the model training process, the following describes the training process of the game item supply and demand change prediction model in detail.

[0147] Step S104: Use the first change data and the second change data as training samples to train the preset initial game item supply and demand change prediction model, and generate a prediction model corresponding to the time category for predicting the game item supply and demand changes in the game scene.

[0148] In an optional embodiment of this application, the supply and demand change model for the props is implemented based on the Holt-Winters (HW) method. The Holt-Winters algorithm is a time series analysis and prediction method applicable to non-stationary sequences with linear trends and fixed periods. It is divided into additive and multiplicative models. In this embodiment, the additive model is used.

[0149] In this embodiment of the application, the initial game item supply and demand change prediction model obtained based on the HW algorithm is used to predict the game item supply and demand change data for another period of time after the first period of time, based on the game item supply and demand change data for a period of time within the game scene. The time range of the first period of time and the time range of the other period of time are the model parameters of the model.

[0150] In step S104 of this application, training the initial game item supply and demand change prediction model is actually finding the most suitable model parameters based on the first change data and the second change data.

[0151] In specific applications, the model parameters are specifically represented by the number of time nodes. For example, assuming the first preset time period is 60 minutes long, the second preset time period is 30 minutes long, and each 5 minutes is a time node, then the parameters of the initial game item supply and demand change prediction model are (12, 6). That is, the game item supply and demand change prediction model is used to predict the game item supply and demand change data for the next 6 time nodes based on the game item supply and demand change data for the first 12 time nodes. It should be noted that, due to the limitations of the HW algorithm, in one optional embodiment of this application, the first time period should contain at least 7 time nodes, and the number of time nodes in the second time period should be less than the number of nodes in the first time node. Furthermore, once the time length corresponding to each time node is determined (for example, 5 minutes), the length corresponding to each time node remains unchanged during model training.

[0152] Specifically, step S104 above includes the following steps S11 and S12:

[0153] Step S11: Input the first change data into the initial game item supply and demand change prediction model to obtain the prediction data of the game item supply and demand change data in the second time period output by the initial game item supply and demand change prediction model; it should be noted that, inputting the first change data into the initial game item supply and demand change prediction model means inputting the first change data from all historical supply and demand change data under the time category into the initial game item supply and demand change prediction model (hereinafter referred to as the initial model).

[0154] For example, suppose the time category is ordinary holidays, which includes the first historical supply and demand change data, the second historical supply and demand change data, the third historical supply and demand change data, ..., the Nth historical supply and demand change data.

[0155] Among them, the first change data for the first time period corresponding to each historical supply and demand change data are X1, X2, X3, ..., X... nTherefore, step S11 above specifically refers to:

[0156] Input X1 into the initial model to obtain the predicted data of the supply and demand changes of game items during the second time period, output by the initial model.

[0157] Input X2 into the initial model to obtain the predicted data of the supply and demand changes of game items during the second time period, output by the initial model.

[0158] Until X n Input the initial model to obtain the predicted data of the supply and demand changes of game items during the second time period, which is output by the initial model.

[0159] Step S12: Adjust the initial game item supply and demand change prediction model according to the second change data and the prediction data to obtain the game item supply and demand change prediction model.

[0160] In this embodiment of the application, step S12 can be understood as adjusting the parameters of the initial game item supply and demand change model based on the second similarity between the predicted data output by the initial game item supply and demand change model and the second change data (the actual value of the game item supply and demand change data in the second time period).

[0161] In one optional embodiment of this application, the similarity between the predicted data and the second changed data can be expressed in the form of root mean square error. Specifically, the root mean square error is obtained by the following formula (2):

[0162]

[0163] Wherein, RMSE represents the root mean square error (i.e., second similarity) between the second changed data and the predicted data. Representing the predicted data, y i This represents the second change data in the i-th historical supply and demand change data under the stated time category.

[0164] Specifically, the smaller the root mean square error calculated based on the above formula, the higher the similarity between the second changed data and the predicted data; conversely, the larger the root mean square error, the lower the similarity between the second changed data and the preset data.

[0165] It should be noted that the above method of determining the similarity between the second changed data and the predicted data by means of root mean square error is only an optional implementation of this application and is not intended to limit the method of calculating the similarity. For example, the similarity between the second changed data and the predicted data can also be obtained by extracting the data features between the second changed data and the predicted data. This application does not limit this.

[0166] Furthermore, if the calculated second similarity between the second change data and the predicted data is greater than the preset second similarity threshold, it indicates that the training accuracy of the prediction model can be met based on the first change data in the first preset time period and the second change data in the second preset time period. In this case, the prediction model for predicting the supply and demand changes of game props in the game scene can be obtained simply by using the first change data as the input data of the prediction model and the second change data as the output data of the initial model.

[0167] If the calculated second similarity between the second changed data and the predicted data is less than or equal to a preset similarity threshold, then further adjustments need to be made to the input and output parameters of the initial model.

[0168] As mentioned earlier, in this embodiment of the application, the parameters of the initial model are specifically the number of time nodes in the first preset time period and the second preset time period. In other words, adjusting the parameters of the initial model is actually adjusting the number of time nodes corresponding to the input data and output data of the initial model.

[0169] For example, assuming the initial model uses 5 minutes as a time node, and the model parameters are adjusted from the initial preset parameters (12, 6) to (13, 7), it means that the first preset time period is adjusted from 60 minutes to 65 minutes, and the second preset time period is adjusted from 30 minutes to 35 minutes.

[0170] After the adjusted parameters are determined, the supply and demand change data corresponding to the first preset time period after the adjustment of the time length will be used as the first change data.

[0171] The supply and demand change data corresponding to the second preset time period after adjusting the time length is used as the second change data;

[0172] Then, the process continues to execute the step of inputting the first change data into the initial game item supply and demand change prediction model to obtain the prediction data of the game item supply and demand change data in the second time period output by the initial game item supply and demand change prediction model. Then, based on the similarity between the prediction data and the second change data, training samples that meet the second similarity threshold are determined.

[0173] The method for generating a game item supply and demand change prediction model provided in this application divides the generation date of the historical supply and demand change data into time categories, and constructs a prediction model corresponding to the time categories based on the change data at different time points in the historical supply and demand change data. This model is used to predict the supply and demand changes of game items in the game scene, thus achieving the prediction of normal supply and demand change data in the game scene. The model obtained by this method can accurately predict the normal supply and demand relationship of game items in the game scene. The supply and demand change prediction data obtained by the model within a preset time period provides a reference for the actual supply and demand change data generated in the game scene, so that game maintenance personnel can monitor abnormal time nodes in the game scene and then perform effective management operations for the abnormal time nodes.

[0174] Another embodiment of this application provides a method for determining abnormal changes in the supply and demand of game items. This method embodiment is mainly used to describe the application of the game item supply and demand change prediction model. Therefore, the description of the game item supply and demand change prediction model involved in this embodiment is merely illustrative. For relevant details, please refer to the description of the previous embodiment; further elaboration is not required here.

[0175] Please refer to Figure 4 , Figure 4 The flowchart of the method for determining abnormal supply and demand of game props provided in this application embodiment includes the following steps S401 to S403.

[0176] The main idea of ​​the method for determining abnormal changes in the supply and demand of game items provided in this application is as follows: input the supply and demand change data of game items within a certain time period in the game scene into the game item supply and demand change prediction model, predict the supply and demand change data of game items within another time period, and if the similarity between the predicted game item supply and demand change data and the actual supply and demand change data of game items within that time period is lower than a preset similarity threshold, then it is considered that the supply and demand change of game items within that time period has become abnormal.

[0177] Step S401: Obtain first supply and demand change data of game items within a first preset time period and second supply and demand change data of game items within a second preset time period, wherein the time category corresponding to the generation date of the first supply and demand change data is the same as the time category corresponding to the generation date of the second supply and demand change data.

[0178] In this embodiment, the first supply and demand change data for the first preset time period is the game item supply and demand change data for a specific time period corresponding to a certain date, obtained from the game database. For example, suppose the game maintenance personnel want to perform anomaly detection on the supply and demand changes of game items on a specific date, then the first preset time period can be the game item supply and demand change data from 12:01 to 13:00 on that day (here referred to as the first supply and demand change data), and the second preset time period can be the game item supply and demand change data from 13:01 to 13:30 on that day (here referred to as the second supply and demand change data).

[0179] Step S402: Input the first supply and demand change data into the game item supply and demand change prediction model corresponding to the time category to obtain the prediction data of the game item supply and demand change data within the preset second time period output by the game item supply and demand change prediction model.

[0180] In this embodiment, the game item supply and demand change prediction model corresponds to the same time category as the generation date of the game item supply and demand change data in the game scene. For example, assuming the time category includes: ordinary weekdays and ordinary weekends, and the generation time of the first supply and demand change data is Saturday, then the specific time of the game item supply and demand change prediction model is a game item supply and demand change prediction model with the time category of ordinary weekends.

[0181] The reason this application categorizes the game item supply and demand prediction model based on time categories is that the supply and demand relationship of game items within the game scenario varies under different time categories. For example, during normal weekdays, the number of players participating in the game is generally small, and most players log in at night. Therefore, during normal weekdays, the supply and demand of game items is characterized by being stable during the day and fluctuating greatly at night. On normal weekends or holidays, the number of players participating in the game is higher than on normal weekdays, and during normal weekends, the number of players participating in the game during the day and at night is basically the same. The overall supply and demand relationship of game items should also be stable.

[0182] In step S402 above, the first supply and demand change data is input into the game item supply and demand change prediction model to predict the supply and demand change data of the second time period, obtain the prediction data, and then compare the prediction data with the second supply and demand change data generated in the second time period. Based on the comparison result, it is determined whether the supply and demand relationship of the second time period is abnormal.

[0183] Step S403: Based on the similarity between the second supply and demand change data and the predicted data, determine the abnormal situation of the supply and demand change of game items in the second time period.

[0184] In one optional embodiment of this application, the number of time nodes of the output data set in the game item supply and demand change prediction model can also be determined, and then the similarity between the second supply and demand change data and the prediction data can be determined based on the data change characteristics at each time node of the second time period and the data change characteristics at each time node of the prediction data.

[0185] Specifically, the process of setting the time node can be referred to the relevant description of step S6 in the previous embodiment of this application, and will not be described here.

[0186] Specifically, assume that the data change characteristics at each node of the second supply and demand change data are as follows: c1, c2, ..., c m The data change characteristics of each node in the predicted data are d1, d2, ..., d... m The similarity between the predicted data and the second supply and demand change data can be expressed as:

[0187]

[0188] After determining the predicted data and the second supply and demand change data, continue to execute the following steps S13 to S15.

[0189] Step S13: Determine whether the similarity between the second supply and demand change data and the preset data is less than a preset similarity threshold;

[0190] Step S14: If the similarity between the second supply and demand change data and the preset data is less than the preset similarity threshold, then it is determined that the supply and demand change of game props in the second time period is abnormal.

[0191] Step S15: If the similarity between the second supply and demand change data and the preset data is greater than or equal to the preset similarity threshold, then it is determined that the supply and demand change of game props in the second time period is normal.

[0192] Another embodiment of this application provides a device for generating a prediction model of supply and demand changes for game items. Since this device embodiment is basically similar to the method for generating a prediction model of supply and demand changes for game items provided in this application, the description is relatively simple. For relevant parts, please refer to the above description of the method in the embodiments. The following description of the device embodiment is merely illustrative.

[0193] Please refer to Figure 5 , Figure 5A schematic diagram of the structure of a device for generating a prediction model of the supply and demand changes of game props provided in another embodiment of this application.

[0194] The device for generating the game item supply and demand change prediction model includes:

[0195] The first acquisition unit 501 is used to acquire historical supply and demand change data of game props in the game scene;

[0196] Determining unit 502 is used to determine the time category corresponding to the generation date of the historical supply and demand change data;

[0197] The second obtaining unit 503 is used to obtain first change data within a first preset time period and second change data within a second preset time period from the historical supply and demand change data, wherein the first preset time period is earlier than the second preset time period.

[0198] The generation unit 504 is used to train a preset initial game item supply and demand change prediction model using the first change data and the second change data as training samples, and to generate a prediction model corresponding to the time category for predicting the game item supply and demand changes in the game scene.

[0199] Optionally, obtaining historical supply and demand data for game items in the game scene includes:

[0200] Obtain data on the changes in the total amount of game items and the transaction data of game items generated by the historical game behavior of game players in the first game server, wherein the generation date of the historical game behavior of the game players is the same as the generation date of the historical supply and demand change data;

[0201] Based on the changes in the total quantity of game items and the transaction data of game items, the historical supply and demand changes of game items within the game scene are determined.

[0202] Optionally, obtaining historical supply and demand data for game items within the game scene includes:

[0203] Obtain data on the changes in the total amount of tradable game items listed on the second game server and data on game item transactions generated by players' game behavior, wherein the generation date of the players' game behavior is the same as the generation date of the historical supply and demand change data;

[0204] Based on the quantity of the tradable game items and the game item trading data, determine the historical supply and demand changes of game items within the game scene.

[0205] Optionally, determining the time category corresponding to the generation date of the historical supply and demand change data includes:

[0206] Obtain other supply and demand change data corresponding to different time categories;

[0207] Obtain the first data change characteristics of other supply and demand change data corresponding to each different time category and the second data change characteristics of the historical supply and demand change data;

[0208] Based on the first similarity between the first data change feature and the second data change feature, the time category corresponding to the generation date of the historical supply and demand change data is determined.

[0209] Optionally, the device is also used for:

[0210] Preset date and corresponding time category;

[0211] The time category includes at least one of the following: regular weekday, regular holiday, game update day.

[0212] Optionally, obtaining the first data change characteristics of other supply and demand change data corresponding to each different time category and the second data change characteristics of the historical supply and demand change data includes:

[0213] The historical supply and demand change data is divided into time nodes according to a preset first time interval, and the supply and demand change characteristics at each time node of the historical supply and demand change data are determined as the first data change characteristics.

[0214] The other supply and demand change data are divided into time nodes according to the first time interval, and the supply and demand change characteristics at each node of the other supply and demand change data are determined as the second data change characteristics.

[0215] Optionally, determining the time category corresponding to the generation date of the historical supply and demand change data based on the similarity between the first data change feature and the second data change feature includes:

[0216] Based on the second similarity between the first data change characteristics and the second data change characteristics at each same time point, the first similarity between the historical supply and demand change data and other supply and demand change data corresponding to each time category is determined respectively;

[0217] The time category corresponding to the other supply and demand change data that exceeds the preset similarity threshold is taken as the time category corresponding to the generation date of the historical supply and demand change data.

[0218] Optionally, the step of using the first change data and the second change data as training samples to train a preset initial game item supply and demand change prediction model, generating a prediction model corresponding to the time category for predicting game item supply and demand changes within the game scene, includes:

[0219] The first change data is input into the initial game item supply and demand change prediction model to obtain the prediction data of the game item supply and demand change data in the second time period output by the initial game item supply and demand change prediction model.

[0220] Based on the second change data and the predicted data, the initial game item supply and demand change prediction model is adjusted to obtain the game item supply and demand change prediction model.

[0221] Optionally, the following condition may be used as the condition for terminating the training of the preset initial game item supply and demand change prediction model: the similarity between the second change data and the prediction data reaches a predetermined threshold.

[0222] Another embodiment of this application provides a device for determining abnormal changes in the supply and demand of game items. Since this device embodiment is basically similar to the method for determining abnormal changes in the supply and demand of game items provided in this application, the description is relatively simple. For relevant parts, please refer to the above description of the method in the embodiments. The following description of the device embodiment is merely illustrative.

[0223] Please refer to Figure 6 , Figure 6 A schematic diagram of a device for determining abnormal supply and demand of game items, provided in another embodiment of this application.

[0224] The device for determining abnormal changes in the supply and demand of game items includes:

[0225] The third obtaining unit 601 is used to obtain first supply and demand change data of game items within a first preset time period and second supply and demand change data of game items within a second preset time period, wherein the time category corresponding to the generation date of the first supply and demand change data is the same as the time category corresponding to the generation date of the second supply and demand change data.

[0226] The model prediction unit 602 is used to input the first supply and demand change data into the game item supply and demand change prediction model corresponding to the time category, and obtain the prediction data of the game item supply and demand change data within a preset second time period output by the game item supply and demand change prediction model.

[0227] The anomaly detection unit 603 is used to determine anomalies in the supply and demand of game items within the second time period based on the similarity between the second supply and demand change data and the predicted data.

[0228] Optionally, determining abnormal situations in the supply and demand of game items within the second time period based on the similarity between the second supply and demand change data and the predicted data includes:

[0229] Determine whether the similarity between the second supply and demand change data and the preset data is less than a preset similarity threshold;

[0230] If the similarity between the second supply and demand change data and the preset data is less than the preset similarity threshold, then it is determined that the supply and demand change of game items in the second time period is abnormal.

[0231] If the similarity between the second supply and demand change data and the preset data is greater than or equal to the preset similarity threshold, then it is determined that the supply and demand change of game items within the second time period is normal.

[0232] Optionally, the device is further configured to: preset the time category corresponding to the date;

[0233] The time category includes at least one of the following: regular weekday, regular holiday, game update day.

[0234] Another embodiment of this application provides an electronic device, please refer to... Figure 7 , Figure 7 This is a schematic diagram of an electronic device structure provided for another embodiment of this application.

[0235] The electronic device includes: a processor 701;

[0236] The memory 702 is used to store the program of the method, which, when read and executed by the processor 501, executes any one of the methods in the above embodiments.

[0237] Another embodiment of this application provides a computer storage medium storing a computer program that, when executed, performs any of the methods described in the above embodiments.

[0238] It should be noted that the detailed description of the electronic device and computer storage medium provided in the embodiments of this application can be found in the relevant description of the above method embodiments provided in this application, and will not be repeated here.

[0239] Although this application discloses preferred embodiments as described above, it is not intended to limit this application. Any person skilled in the art can make possible changes and modifications without departing from the spirit and scope of this invention. Therefore, the scope of protection of this application should be determined by the scope defined in the claims of this application.

[0240] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0241] Memory may include non-persistent storage in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0242] 1. Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information by any method or technology. Information can be computer-readable instructions, data structures, modules of programs, 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 cassettes, magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include non-transitory computer-readable media, such as modulated data signals and carrier waves.

[0243] 2. Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

Claims

1. A method for generating a prediction model of supply and demand changes for game items, characterized in that, include: Obtain historical supply and demand change data for at least one game item in the game scene, wherein the historical supply and demand change data corresponds to the generation date information of the historical supply and demand change data; Determine the time category corresponding to the generation date of the historical supply and demand change data; From the historical supply and demand change data, obtain first change data within a first preset time period and second change data within a second preset time period, wherein the first preset time period is earlier than the second preset time period; The first change data and the second change data are used as training samples to train the preset initial game item supply and demand change prediction model, thereby generating a prediction model corresponding to the time category for predicting game item supply and demand changes in the game scene. The step of using the first change data and the second change data as training samples to train a preset initial game item supply and demand change prediction model includes: Input the first change data from all historical supply and demand change data under the time category into the initial game item supply and demand change prediction model; Determining the time category corresponding to the generation date of the historical supply and demand change data includes: Obtain other supply and demand change data corresponding to different time categories; Obtain the first data change characteristics of other supply and demand change data corresponding to each different time category and the second data change characteristics of the historical supply and demand change data; Based on the first similarity between the first data change characteristics and the second data change characteristics, determine the time category corresponding to the generation date of the historical supply and demand change data; The first data change feature for obtaining the other supply and demand change data corresponding to each different time category and the second data change feature for the historical supply and demand change data include: The historical supply and demand change data is divided into time nodes according to a preset first time interval, and the supply and demand change characteristics at each time node of the historical supply and demand change data are determined as the first data change characteristics. The other supply and demand change data are divided into time nodes according to the first time interval, and the supply and demand change characteristics at each node of the other supply and demand change data are determined as the second data change characteristics.

2. The method according to claim 1, characterized in that, The acquisition of historical supply and demand data for game items in the game scene includes: Obtain data on the changes in the total amount of game items and the transaction data of game items generated by the historical game behavior of game players in the first game server, wherein the generation date of the historical game behavior of the game players is the same as the generation date of the historical supply and demand change data; Based on the changes in the total quantity of game items and the transaction data of game items, the historical supply and demand changes of game items within the game scene are determined.

3. The method according to claim 1, characterized in that, The acquisition of historical supply and demand data for game items within the game scene includes: Obtain data on the changes in the total amount of tradable game items listed on the second game server and data on game item transactions generated by players' game behavior, wherein the generation date of the players' game behavior is the same as the generation date of the historical supply and demand change data; Based on the quantity of the tradable game items and the game item trading data, determine the historical supply and demand changes of game items within the game scene.

4. The method according to claim 1, characterized in that, The method further includes: Preset date and corresponding time category; The time category includes at least one of the following: regular weekday, regular holiday, game update day.

5. The method according to claim 1, characterized in that, The step of determining the time category corresponding to the generation date of the historical supply and demand change data based on the similarity between the first data change feature and the second data change feature includes: Based on the second similarity between the first data change characteristics and the second data change characteristics at each same time point, the first similarity between the historical supply and demand change data and other supply and demand change data corresponding to each time category is determined respectively; The time category corresponding to the other supply and demand change data whose first similarity is greater than the preset first similarity threshold is taken as the time category corresponding to the generation date of the historical supply and demand change data.

6. The method according to claim 1, characterized in that, Obtaining first change data within a first preset time period and second change data within a second preset time period from the historical supply and demand change data includes: Obtain the first preset time period and the second preset time period set for the initial game item supply and demand change prediction model; Extract the first change data corresponding to the first preset time period from the historical supply and demand change data; Extract the second change data corresponding to the second preset time period from the historical supply and demand change data.

7. The method according to claim 1, characterized in that, The step of using the first change data and the second change data as training samples to train a preset initial game item supply and demand change prediction model, generating a prediction model corresponding to the time category for predicting game item supply and demand changes within the game scene, includes: The first change data is input into the initial game item supply and demand change prediction model to obtain the first prediction data of the game item supply and demand change data within the second preset time period output by the initial game item supply and demand change prediction model. If the second similarity between the second change data and the first prediction data is greater than a preset second similarity threshold, the initial game item supply and demand change prediction model is trained based on the first change data and the second change data to obtain the game item supply and demand change prediction model.

8. The method according to claim 7, characterized in that, The method further includes: If the second similarity between the second changed data and the first predicted data is less than or equal to a preset second similarity threshold, the duration of the first preset time period and the second preset time period is adjusted according to the adjustment range of the first preset time period and the second preset time period. The supply and demand change data corresponding to the first preset time period after adjusting the time length is used as the first change data, and the supply and demand change data corresponding to the second preset time period after adjusting the time length is used as the second change data. Then, the process returns to the step of inputting the first change data into the initial game item supply and demand change prediction model to obtain the prediction data of the game item supply and demand change data in the second preset time period output by the initial game item supply and demand change prediction model.

9. A method for determining abnormal changes in the supply and demand of game items, characterized in that, include: Obtain first supply and demand change data of game items within a first preset time period and second supply and demand change data of game items within a second preset time period, wherein the time category corresponding to the generation date of the first supply and demand change data is the same as the time category corresponding to the generation date of the second supply and demand change data. The first supply and demand change data is input into the game item supply and demand change prediction model corresponding to the time category to obtain the predicted data of game item supply and demand change data within the preset second time period output by the game item supply and demand change prediction model. Based on the similarity between the second supply and demand change data and the predicted data, anomalies in the supply and demand changes of game items within the second time period are determined. The step of determining abnormal situations in the supply and demand of game items within the second time period based on the similarity between the second supply and demand change data and the predicted data includes: Determine whether the similarity between the second supply and demand change data and the predicted data is less than a preset similarity threshold; If the similarity between the second supply and demand change data and the predicted data is less than a preset similarity threshold, then the supply and demand change of game items in the second time period is determined to be abnormal. If the similarity between the second supply and demand change data and the predicted data is greater than or equal to a preset similarity threshold, then the supply and demand change of game items within the second time period is determined to be normal.

10. The method according to claim 9, characterized in that, The method further includes: Preset date and corresponding time category; The time category includes at least one of the following: regular weekday, regular holiday, game update day.

11. A device for generating a prediction model of supply and demand changes for game props, characterized in that, include: The first acquisition unit is used to acquire historical supply and demand change data of game items in the game scene; A determining unit is used to determine the time category corresponding to the generation date of the historical supply and demand change data; The second obtaining unit is used to obtain first change data within a first preset time period and second change data within a second preset time period from the historical supply and demand change data, wherein the first preset time period is earlier than the second preset time period. The generation unit is used to train a preset initial game item supply and demand change prediction model using the first change data and the second change data as training samples, and to generate a prediction model corresponding to the time category for predicting the game item supply and demand change in the game scene. The step of using the first change data and the second change data as training samples to train a preset initial game item supply and demand change prediction model includes: Input the first change data from all historical supply and demand change data under the time category into the initial game item supply and demand change prediction model; Determining the time category corresponding to the generation date of the historical supply and demand change data includes: Obtain other supply and demand change data corresponding to different time categories; Obtain the first data change characteristics of other supply and demand change data corresponding to each different time category and the second data change characteristics of the historical supply and demand change data; Based on the first similarity between the first data change characteristics and the second data change characteristics, determine the time category corresponding to the generation date of the historical supply and demand change data; The first data change feature for obtaining the other supply and demand change data corresponding to each different time category and the second data change feature for the historical supply and demand change data include: The historical supply and demand change data is divided into time nodes according to a preset first time interval, and the supply and demand change characteristics at each time node of the historical supply and demand change data are determined as the first data change characteristics. The other supply and demand change data are divided into time nodes according to the first time interval, and the supply and demand change characteristics at each node of the other supply and demand change data are determined as the second data change characteristics.

12. A device for determining abnormal changes in the supply and demand of game items, characterized in that, include: The third obtaining unit is used to obtain first supply and demand change data of game items within a first preset time period and second supply and demand change data of game items within a second preset time period, wherein the time category corresponding to the generation date of the first supply and demand change data is the same as the time category corresponding to the generation date of the second supply and demand change data. The model prediction unit is used to input the first supply and demand change data into the game item supply and demand change prediction model corresponding to the time category, and obtain the prediction data of the game item supply and demand change data within a preset second time period output by the game item supply and demand change prediction model. An anomaly detection unit is used to determine anomalies in the supply and demand of game items within a second time period based on the similarity between the second supply and demand change data and the predicted data. The step of determining abnormal situations in the supply and demand of game items within the second time period based on the similarity between the second supply and demand change data and the predicted data includes: Determine whether the similarity between the second supply and demand change data and the predicted data is less than a preset similarity threshold; If the similarity between the second supply and demand change data and the predicted data is less than a preset similarity threshold, then the supply and demand change of game items in the second time period is determined to be abnormal. If the similarity between the second supply and demand change data and the predicted data is greater than or equal to a preset similarity threshold, then the supply and demand change of game items within the second time period is determined to be normal.

13. An electronic device, characterized in that, include: processor; A memory for storing a program of a method, which, when read and executed by a processor, performs the method according to any one of claims 1-10.

14. A computer storage medium, characterized in that, The computer storage medium stores a computer program, which, when executed, performs the method described in any one of claims 1-10.

Images