Game account transaction probability prediction method and system
By cleaning and extracting key assets from game accounts to form feature vectors, calculating similarity, and using an accelerated failure model to predict transaction probability, this technology solves the problem of inaccurate prediction of game account transaction probability in existing technologies, achieving more accurate prediction and higher transaction trust.
Patent Information
- Application Number
- CN202511520814.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-10-23
- Publication Date
- 2026-02-17
AI Technical Summary
Existing technologies cannot accurately predict the probability of game account transactions, especially in environments with information asymmetry, making it difficult to provide effective decision support for both buyers and sellers. Furthermore, traditional methods cannot fully reflect the complexity of game account transactions.
By acquiring the raw data of the target game account, cleaning and structuring it, extracting key assets to form feature vectors, calculating similarity, filtering similar game accounts, and using an accelerated failure model to predict the probability of a transaction.
It improves the accuracy of predicting the probability of game account transactions, enhances the transparency and understandability of the transaction process, reduces transaction risks, and promotes the healthy development of the market and the establishment of trust.
Smart Images
Figure CN121544307A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application relates to the technical field of data processing, in particular to a game account transaction probability prediction method and system. BACKGROUND
[0002] With the vigorous development of the network game industry, game accounts as a special virtual asset. More and more players and investors are willing to buy and sell game accounts to realize the value-added and realization of their virtual assets. This market not only attracts a large number of buyers and sellers, but also promotes the emergence of corresponding trading platforms and intermediary services. The transaction of game accounts is full of uncertainty, and the buyers and sellers face many challenges. The seller hopes to realize the value quickly at a reasonable price, but due to the information asymmetry in the market, how to price to ensure the best interests in market competition has always been a difficult problem. The buyer hopes to buy a game account with high cost performance and meet the demand, and needs to evaluate the value and transaction period of the account. Therefore, how to accurately predict the transaction probability and transaction period of a game account has become a problem that needs to be solved by buyers and sellers and trading platforms.
[0003] Currently, the prediction method in the game account transaction market, especially the prediction of game account transaction probability, is still in its infancy. Most trading platforms rely on traditional market analysis methods and experience to estimate prices and predict transactions.
[0004] However, the current game account transaction probability prediction method often cannot provide accurate and scientific prediction, especially in an information asymmetry environment, it is difficult to provide effective decision support for buyers and sellers. In addition, due to the high individualization and diversity of game accounts, traditional prediction methods based on single factors such as price and game type cannot fully reflect the complexity of account transactions in the market. SUMMARY
[0005] In view of the above deficiencies of the prior art, the purpose of the embodiments of the present application is to provide a game account transaction probability prediction method, which can solve the technical problems that the current game account transaction probability prediction method often cannot provide accurate and scientific prediction, especially in an information asymmetry environment, it is difficult to provide effective decision support for buyers and sellers, and due to the high individualization and diversity of game accounts, traditional prediction methods based on single factors such as price and game type cannot fully reflect the complexity of account transactions in the market.
[0006] The first aspect of the embodiments of the present application proposes a game account transaction probability prediction method, comprising: S1: obtaining original data of a target game account; S2: cleaning the original data to obtain structured data; S3: extracting key assets of the target game account from the structured data to obtain a feature vector; S4: calculating similarity between the target game account and other game accounts according to the feature vector; S5: screening game accounts with similarity greater than a preset similarity from other game accounts as similar game accounts; S6: predicting a transaction probability of the target game account through an accelerated failure model according to transaction data of the similar game accounts.
[0007] The second aspect of the embodiment of the application provides a game account transaction probability prediction system, comprising a processor and a memory. The memory stores programs or instructions executable on the processor, and the programs or instructions are executed by the processor to implement the steps of the game account transaction probability prediction method according to the first aspect.
[0008] The third aspect of the embodiment of the application provides a readable storage medium, and the readable storage medium stores programs or instructions, and the programs or instructions are executed by a processor to implement the steps of the game account transaction probability prediction method according to the first aspect.
[0009] The technical scheme provided by the embodiment of the application has at least the following beneficial effects: In the embodiment of the application, similar game accounts are screened out through accurate similarity matching, the high correlation between the similar game accounts in the game account database and the target game account is ensured, the transaction probability of the target game account is predicted through an accelerated failure model, the prediction accuracy of the game account transaction probability is improved, the complexity of account transactions in the market can be fully reflected, and effective decision support can be provided for both buyers and sellers. BRIEF DESCRIPTION OF DRAWINGS
[0010] The accompanying drawings are only for the purpose of illustrating specific embodiments and are not considered as limiting the application. Throughout the drawings, the same reference signs indicate the same components. Obviously, the accompanying drawings described below are only some embodiments described in the embodiments of the application, and other drawings can be obtained by those skilled in the art without creative effort.
[0011] Figure 1 is a flowchart of a game account transaction probability prediction method provided by the embodiment of the application; Figure 2 is a structural diagram of a game account transaction probability prediction system provided by the embodiment of the application. DETAILED DESCRIPTION
[0012] In order to make the person skilled in the art better understand the technical solutions in the embodiments of the present application, the technical solutions of the present application will be described clearly and completely in conjunction with the drawings. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. It should be understood that these descriptions are only exemplary and are not used to limit the scope of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without creative labor should belong to the scope of protection of the present application.
[0013] The game account transaction probability prediction method provided by the embodiments of the present application will be described in detail below in conjunction with the drawings, specific embodiments and application scenarios.
[0014] Referring to the drawings, the flowchart of the game account transaction probability prediction method provided by the embodiments of the present application is shown. Figure 1
[0015] The game account transaction probability prediction method provided by the embodiments of the present application can include the following steps: S1: Obtain the original data of the target game account.
[0016] The original data is generally unstructured data such as game account description text and game asset screenshots. Directly using the original data not only has data quality problems (part of the data is provided by the seller and may not be real), but also has high requirements for model complexity (large model). Therefore, data cleaning work is usually required.
[0017] S2: Clean the original data to obtain structured data.
[0018] In one possible implementation, after S2 and before S3, there is also: judging whether the data quality of the structured data has a problem. If yes, return to S1. Otherwise, proceed to the next step.
[0019] In the present application, there are mainly three data quality problems: data attribution, business attribution and algorithm attribution. Data attribution refers to asset missing, asset quantity exceeding the upper and lower limits, etc. Business attribution refers to game account description text being too short (missing valuable information), conjoined numbers (multiple numbers being sold together), etc. Algorithm attribution refers to suspected high-priced transactions (low-priced high sales), suspected low-priced transactions (high-priced low sales), etc.
[0020] The cleaning process generally needs to structure the original data and clean it according to the official game asset data. For example, there are N+M assets of a certain type A in the game asset description text, but only N assets in the official data. For this type of data, filtering is required, etc.
[0021] In the embodiments of the present application, by checking and cleaning potential problems in the data (such as asset missing, incomplete description, abnormal price, etc.), inconsistent or unrealistic data can be eliminated to avoid negative impact on subsequent analysis and prediction model. By comparison and filtering with official data, the consistency and reliability of the data can be ensured, thereby improving the accuracy and credibility of the prediction results, and providing more valuable reference and decision support for the trading platform and users.
[0022] S3: Extracting key assets of the target game account from the structured data to obtain a feature vector.
[0023] It should be noted that the key assets are different in different games, so this part needs to be personalized configured. After the game account data cleaning is completed, the key assets in each game account are extracted as a feature vector according to the game key asset configuration, and the feature vector is stored in the vector library.
[0024] Specifically, the original data (game assets, account attributes, generally little change) of each game account is compressed to obtain a feature vector (an abstract representation of the game account).
[0025] In the embodiments of the present application, the original data of each game account is compressed into a feature vector and stored in the vector library, which can effectively convert complex game account information into a concise and representative abstract representation. The advantage of this method is that it can extract unique key assets in different games and perform personalized configuration for each game account, thereby facilitating subsequent similarity calculation and transaction probability prediction. Through the compression of the feature vector, the complexity of data processing is reduced, and important information of the game account is retained, so that the model is more efficient in processing, and the accuracy and operability of the prediction are improved.
[0026] In one possible implementation, after S3 and before S4, it further includes: judging whether the feature vector of the target game account already exists in the vector library. If yes, go to the next step. Otherwise, connect the game account database, clean the data, extract the key information, and build the vector library.
[0027] In the embodiments of the present application, the step of judging whether the feature vector of the target game account already exists in the vector library can effectively avoid repeated calculation and improve efficiency. When the feature vector already exists, the process of re-computation can be directly skipped, saving computing resources and time. If the feature vector does not exist, the data is cleaned, the key information is extracted, and the vector library is built by connecting the game account database, which can ensure the integrity and timeliness of the vector library, update and expand the data in time, and ensure the accuracy and prediction ability of the model. Through this way of dynamically managing the vector library, the system running efficiency is improved, and the latest and accurate data is ensured.
[0028] S4: Calculate the similarity between the target game account and other game accounts based on the feature vector.
[0029] It should be noted that since the application market of this invention is the second-hand trading of game accounts, stronger explanatory power is required to reassure sellers and buyers about the transaction. Therefore, this invention adopts a similarity-based probability prediction method.
[0030] In one possible implementation, S4 specifically involves: calculating the similarity between the target game account and other game accounts based on the feature vectors, using cosine similarity or Euclidean distance.
[0031] Cosine similarity measures the similarity between two vectors by calculating the cosine of the angle between them. The value ranges from -1 to 1; a value closer to 1 indicates greater similarity. It is suitable for measuring similarity along vector directions, ignoring the magnitude of the vectors. It is commonly used in text mining and recommendation systems.
[0032] Euclidean distance, on the other hand, calculates the straight-line distance between two vectors. The smaller the value, the closer the two vectors are. It is suitable for measuring the actual distance between vectors in space.
[0033] Furthermore, cosine similarity and Euclidean distance have different advantages in different application scenarios. Cosine similarity focuses more on comparing directional similarity, while Euclidean distance emphasizes the actual numerical difference.
[0034] It should be noted that by using precise similarity matching, we can ensure a high degree of correlation between the game accounts in the game account database and the target game accounts, thereby improving the accuracy of predicting the probability of a successful transaction.
[0035] In this embodiment of the invention, a similarity-based transaction probability prediction method is employed, which provides stronger interpretability and helps sellers and buyers better understand the prediction results. By calculating the similarity between the target game account and other accounts, sellers can clearly see the transaction history of other accounts similar to their own, thereby making more reasonable pricing and sales decisions; buyers, on the other hand, can understand the potential value of the target account based on the historical transaction information of similar accounts, increasing their purchasing confidence. This method not only enhances the transparency and understandability of the transaction process but also reduces the risks for both parties, promoting the healthy development of the market and the establishment of trust.
[0036] S5: Select game accounts with a similarity greater than the preset similarity from other game accounts and use them as similar game accounts.
[0037] Those skilled in the art can set the preset similarity level according to the actual situation, and the present invention does not limit it.
[0038] Furthermore, it can provide transaction information for similar game accounts, allowing sellers and buyers to trade with confidence.
[0039] S6: Based on the transaction data of similar game accounts, predict the transaction probability of the target game account through an accelerated failure model.
[0040] The Accelerated Failure Time (AFT) model is a statistical model used for survival analysis, primarily for predicting the timing of events. In the AFT model, it is assumed that different factors can accelerate or delay the occurrence of an event. The model predicts the survival time of the target object by multiplying the effects of these influencing factors by the survival time. Specifically, the AFT model uses a logarithmic transformation of the survival time, a weighted sum of covariates to describe the changes in survival time, and employs maximum likelihood estimation to fit the data, thereby effectively predicting the probability and timing of event occurrence. This model has strong flexibility and interpretability, and is applicable to prediction scenarios involving various time-delayed or periodic events.
[0041] It should be noted that, considering the game account trading market, in addition to the game account itself (popular game assets, game account value, etc.) affecting its transaction probability, dynamic information such as exposure frequency, listing price, and seller response speed also significantly influence the transaction probability. Therefore, this is a similarity measurement scenario with dynamic contextual information (covariates), and thus this patent uses an accelerated failure model to predict the transaction probability of game accounts.
[0042] Specifically, given a target game account, the feature vector of the target game account is first obtained. Then, similar game accounts of the target game account are obtained from the vector library. Finally, the AFT model is used to predict the transaction probability of the target game account by combining the transaction information (transaction duration, order price, etc.) of similar game accounts.
[0043] In one possible implementation, S6 specifically includes sub-steps S601 to S604: S601: Using data from similar game accounts, fit an accelerated failure model: in, This indicates the survival time, which is the time from the start of the transaction to the completion of the game account sale. Represents the intercept parameter. Indicates system parameters, Representing covariates, Indicates the scale parameter. This represents the random error term.
[0044] It should be noted that by introducing covariates and integrating order price factors, the impact of market prices on the probability of a transaction can be reflected more accurately.
[0045] S602: Given the covariates corresponding to the target game account, calculate the survival function value: in, Indicates that given covariates Under the conditions t The probability of survival after a certain point in time. Represents the baseline survival function. t Indicates time, Representing covariates, This represents an exponential function with the natural constant as its base.
[0046] It should be noted that conditional probability prediction is provided, and the prediction results can be dynamically adjusted based on the account's listing time.
[0047] S603: Using the fitted accelerated failure model, calculate the survival probability given the covariates corresponding to the target game account: in, Represents conditional probability. Indicates the start time. Indicates the length of the future time window. Indicates that given covariates Under these conditions, survival time is Time to The probability between time points.
[0048] S604: Based on survival probability, predict the probability of a target game account being traded within a specific time period.
[0049] In this embodiment of the invention, by using transaction data of similar game accounts and combining it with an Accelerated Failure Model (AFT model) to predict the transaction probability of a target game account, the market performance of the target account can be estimated more accurately. The AFT model comprehensively considers multiple factors, such as price and game assets, to predict the timing and likelihood of an account's transaction, thereby providing sellers with a reasonable pricing basis and helping them optimize their sales strategies. Simultaneously, buyers can make more informed purchasing decisions based on the predicted transaction probability, reducing risk. This method not only improves the accuracy and reliability of predictions but also provides strong data support for both buyers and sellers, enhancing the transparency and trust in market transactions.
[0050] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following: In this embodiment of the invention, by combining a rule engine and an AI model for dual review, the efficiency and accuracy of the review process can be improved while ensuring prescription compliance. Rigid review uses the rule engine to check for obvious medication problems in real time, while flexible review utilizes the AI model to handle complex, personalized situations, ensuring that prescriptions better meet the specific needs of patients. This not only identifies potential risks in real time, avoiding the lag of post-event review, but also reduces errors in manual review, improving the comprehensiveness and flexibility of prescription review, thereby enhancing medication safety and the quality of medical care.
[0051] Reference manual attached Figure 2 The diagram shows a structural schematic of a game account transaction probability prediction system provided by an embodiment of the present invention.
[0052] This invention provides a game account transaction probability prediction system 20, comprising: The acquisition module 201 is used to acquire the raw data of the target game account.
[0053] The cleaning module 202 is used to clean the raw data to obtain structured data.
[0054] Extraction module 203 is used to extract key assets of the target game account from structured data to obtain feature vectors.
[0055] The calculation module 204 is used to calculate the similarity between the target game account and other game accounts based on the feature vector.
[0056] The filtering module 205 is used to filter out game accounts with a similarity greater than a preset similarity from other game accounts and to identify them as similar game accounts.
[0057] The prediction module 206 is used to predict the probability of a target game account being traded based on the transaction data of similar game accounts and by using an accelerated failure model.
[0058] In one possible implementation, the game account transaction probability prediction system further includes: The first judgment module is used to determine whether there are any issues with the quality of the structured data. If so, the retrieval module is called. Otherwise, proceed to the next step.
[0059] In one possible implementation, the game account transaction probability prediction system further includes: The second judgment module is used to determine whether the feature vector of the target game account already exists in the vector library. If yes, proceed to the next step. Otherwise, connect to the game account database, clean the data, extract key information, and build the vector library.
[0060] In one possible implementation, the computing module 204 is specifically used for: Based on the feature vectors, cosine similarity or Euclidean distance is used to calculate the similarity between the target game account and other game accounts.
[0061] In one possible implementation, the prediction module 206 is specifically used for: Using data from similar game accounts, fit an accelerated failure model: in, This indicates the survival time, which is the time from the start of the transaction to the completion of the game account sale. Represents the intercept parameter. Indicates system parameters, Representing covariates, Indicates the scale parameter. This represents the random error term.
[0062] Given the covariates corresponding to the target game account, calculate the survival function value: in, Indicates that given covariates Under the conditions t The probability of survival after a certain point in time. Represents the baseline survival function. t Indicates time, Representing covariates, This represents an exponential function with the natural constant as its base.
[0063] By fitting the completed accelerated failure model, and given the covariates corresponding to the target game account, the survival probability is calculated: in, Represents conditional probability. Indicates the start time. Indicates the length of the future time window. Indicates that given covariates Under these conditions, survival time is Time to The probability between time points.
[0064] Based on survival probability, predict the probability of a target game account being successfully traded within a specific time period.
[0065] The game account transaction probability prediction system 20 provided by the present invention can implement the steps of the above-mentioned game account transaction probability prediction method and achieve the same technical effect. To avoid repetition, the present invention will not elaborate further.
[0066] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the embodiments of the present invention, and are not intended to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the protection scope of the present invention.
Claims
1. A method for predicting the probability of a game account transaction, characterized in that, include: S1: Obtain the raw data of the target game account; S2: Clean the raw data to obtain structured data; S3: Extract the key assets of the target game account from the structured data to obtain a feature vector; S4: Calculate the similarity between the target game account and other game accounts based on the feature vector; S5: Select game accounts with a similarity greater than the preset similarity from other game accounts and use them as similar game accounts; S6: Based on the transaction data of the similar game accounts, predict the transaction probability of the target game account using the accelerated failure model.
2. The method for predicting the probability of a game account transaction according to claim 1, characterized in that, After S2 and before S3, it also includes: Determine if there is a problem with the quality of the structured data; if so, return to S1; otherwise, proceed to the next step.
3. The method for predicting the probability of game account transactions according to claim 1, characterized in that, After S3 and before S4, it also includes: Determine whether the feature vector of the target game account already exists in the vector library; if yes, proceed to the next step; otherwise, connect to the game account database, clean the data, extract key information, and construct the vector library.
4. The method for predicting the probability of game account transactions according to claim 1, characterized in that, Specifically, S4 is: Based on the feature vector, the similarity between the target game account and other game accounts is calculated using cosine similarity or Euclidean distance.
5. The method for predicting the probability of a game account transaction according to claim 1, characterized in that, S6 specifically includes: S601: Using the data from the similar game accounts, fit the acceleration failure model: in, This indicates the survival time, which is the time from the start of the transaction to the completion of the game account sale. Represents the intercept parameter. Indicates system parameters, Representing covariates, Indicates the scale parameter. Represents the random error term; S602: Given the covariates corresponding to the target game account, calculate the survival function value: in, Indicates that given covariates Under the conditions t The probability of survival after a certain point in time. Represents the baseline survival function. t Indicates time, Representing covariates, Represents an exponential function with the natural constant as its base; S603: Using the fitted accelerated failure model, calculate the survival probability given the covariates corresponding to the target game account: in, Represents conditional probability. Indicates the start time. Indicates the length of the future time window. Indicates that given covariates Under these conditions, survival time is Time to The probability between time points; S604: Based on the survival probability, predict the probability of the target game account being successfully traded within a specific time period.
6. A game account transaction probability prediction system, characterized in that, include: The acquisition module is used to obtain the raw data of the target game account; The cleaning module is used to clean the raw data to obtain structured data; The extraction module is used to extract the key assets of the target game account from the structured data to obtain a feature vector; The calculation module is used to calculate the similarity between the target game account and other game accounts based on the feature vector. The filtering module is used to filter out game accounts with a similarity greater than a preset similarity from other game accounts, and then identify them as similar game accounts. The prediction module is used to predict the probability of a transaction for the target game account based on the transaction data of the similar game accounts and through an accelerated failure model.
7. The game account transaction probability prediction system according to claim 6, characterized in that, Also includes: The first judgment module is used to determine whether there is a problem with the data quality of the structured data; if so, the acquisition module is invoked. Otherwise, proceed to the next step.
8. The game account transaction probability prediction system according to claim 6, characterized in that, Also includes: The second judgment module is used to determine whether the feature vector of the target game account already exists in the vector library; If so, proceed to the next step; Otherwise, connect to the game account database, clean the data, extract key information, and construct the vector library.
9. The game account transaction probability prediction system according to claim 6, characterized in that, The calculation module is specifically used for: Based on the feature vector, the similarity between the target game account and other game accounts is calculated using cosine similarity or Euclidean distance.
10. The game account transaction probability prediction system according to claim 6, characterized in that, The prediction module is specifically used for: Using the data from the similar game accounts, fit the acceleration failure model: in, This indicates the survival time, which is the time from the start of the transaction to the completion of the game account sale. Represents the intercept parameter. Indicates system parameters, Representing covariates, Indicates the scale parameter. Represents the random error term; Given the covariates corresponding to the target game account, calculate the survival function value: Indicates that given covariates Under the conditions t The probability of survival after a certain point in time. Represents the baseline survival function. t Indicates time, Representing covariates, Represents an exponential function with the natural constant as its base; By fitting the accelerated failure model, and given the covariates corresponding to the target game account, the survival probability is calculated: in, Represents conditional probability. Indicates the start time. Indicates the length of the future time window. Indicates that given covariates Under these conditions, survival time is Time to The probability between time points; Based on the survival probability, the probability of the target game account being successfully traded within a specific time period is predicted.