A method for ticket revenue clearing based on multi-element heterogeneous data

Abstract

The application discloses a kind of based on multi-element heterogeneous data ticket revenue clearing, it is related to ticket system share method field.The based on multi-element heterogeneous data ticket revenue clearing, by establishing clearing rule model, the contribution content of each participant is designed as contribution factor, it is realized in the premise of the basis sharing proportion agreed in contract, according to the performance of each participant in its responsible contribution content, through data normalization processing, to form the specific parameterized data that can be calculated, then according to data content calculation, to express the contribution factor of this participant by data assignment, then according to assignment in basis sharing proportion is floated, to mobilize the enthusiasm of participant, bring better performance in its responsible contribution content, while in its responsible contribution content, when the mistake that appears in the contribution content leads to revenue reduction, it is punished and is held accountable, to form the good promotion effect.

Description

Technical Field

[0001] The present invention relates to the technical field of ticketing system revenue sharing methods, and in particular to a method for clearing ticketing revenue based on multi-dimensional heterogeneous data. Background Art

[0002] Nowadays, with the improvement of people's living standards, many regions sell tickets through ticketing systems when holding entertainment activities. After using the ticketing system to count the income, each participant will divide the total income.

[0003] However, the existing revenue settlement is usually based on limited contractual constraints, and is distributed proportionally after deducting costs according to the contractual constraints. However, this distribution method uses a fixed ratio for the profit share of each participant during the distribution. This means that during the event, for participants with lower shares, their participation enthusiasm is low. It is very likely that the participant will cut corners in the part they are responsible for, resulting in a decrease in the effectiveness of the event and ticket revenue. In addition, it is impossible to hold the participant accountable and punish the participant based on responsibility when the total revenue decreases due to mistakes in the part they are responsible for, making the entire distribution plan unreasonable. For this purpose, a method for clearing ticket revenue based on multivariate heterogeneous data is provided. Summary of the Invention

[0004] In response to the shortcomings of the existing technology, the present invention provides a method for clearing ticket revenue based on multi-dimensional heterogeneous data, which solves the problem that the existing revenue clearing is usually based on limited contract constraints, adopts a fixed ratio for the profit sharing of each participant, and cannot hold the participant accountable and punish the participant based on responsibility when the total revenue decreases due to mistakes in the part responsible for the participant, which makes the entire distribution plan unreasonable.

[0005] To achieve the above objectives, the present invention is implemented through the following technical solutions: a method for clearing ticket revenue based on multi-dimensional heterogeneous data, comprising the following steps:

[0006] S1. Data collection: lock users through the ticketing system and obtain multi-dimensional data information about users;

[0007] S2. Data preprocessing and feature normalization: By designing a multivariate data pipeline, we clean the ticketing information-related data generated by users and then extract executable data features from each type of data.

[0008] S3. Establish a clearing rule model, the formula of which is:

[0009]

[0010] in Indicates the final share ratio of participant p; represents the base share ratio of the participant p, which is determined by the pre-contract constraints of the participants; represents the channel contribution factor; represents the external influence factor; represents the attendance rate factor;

[0011] Then, according to the share ratio of the participant p is substituted into the single transaction share calculation, the total profit share obtained by the participant in the i th transaction is calculated, and the formula can be expressed as:

[0012]

[0013] wherein represents the net profit of a single transaction; represents the fraud transaction filtering;

[0014] S4, clearing execution and verification optimization, by bringing the actual ticket revenue and related data into the model calculation to obtain the specific value of the profit share of each participant, then according to the specific data and the expected share ratio of the participant, it is verified whether it is reasonable, and the model parameter optimization is carried out according to the analysis result.

[0015] Preferably, in the data collection process, by identifying and accessing multiple data sources, including ticketing systems and transaction systems of various distribution platforms, transaction data is directly obtained, then user information in the transaction data is used to analyze the user, to obtain user behavior data and user portrait data, combined with on-site equipment to obtain actual use data, external environment data and text feedback data are obtained through the external Internet, contract clause data is input manually, mainly the value of the base share ratio of each participant , then by designing a cost calculation model, analyzing cost data, to calculate .

[0016] Preferably, in step S2, in the data cleaning process, reusable cleaning rules and pipelines are established, and then the data is processed, which includes;

[0017] Structured data, processing missing values, outlier detection and processing, data type conversion, unique identifier generation and matching;

[0018] Semi-structured data, parse ticket order data, extract key fields and structure.

[0019] Unstructured text data, text cleaning, extracting keywords in the text and normalizing positive or negative according to the keywords;

[0020] Time series data, processing timestamp formats, aligning time series, and processing out-of-order data;

[0021] Spatial data, standardize geographic coordinates, and calculate the distance between the ticket purchase location and the venue.

[0022] Preferably, in said S3, the channel contribution factor Normalized data processing is performed to obtain the calculation model of the channel contribution factor, and its formula is:

[0023]

[0024] in represents the quantitative value of the contribution of channel c, represents the normalized value of the contribution ratio of channel c, the weight α=03 represents the control parameter of the reward and punishment range, and b=05 is the expected contribution ratio threshold; Represents the final channel contribution factor of the participants related to channel c.

[0025] Preferably, in said S3, the external influencing factors Perform normalization processing to obtain its calculation model, and its formula is:

[0026]

[0027] in , represents the penalty coefficient for the responsible party, Represents the negative sentiment index, which belongs to the negative sentiment data in unstructured text data, including the sentiment evaluation of the ticketing system, parties, and environmental factors, and is normalized according to positive or negative; , represents the risk sensitivity coefficient; An exponential decay function representing the negative sentiment index.

[0028] Preferably, the attendance factor The model calculation formula is:

[0029]

[0030] Where r represents the actual attendance rate; Indicates the minimum share ratio. is the threshold parameter, trigger thresholds for penalties; The threshold for full share.

[0031] Preferably, the net income from a single transaction When calculating, by introducing the cost calculation model, the formula is expressed as;

[0032]

[0033] in represents the total cost borne by participant p; Indicates the single transaction fee; Indicates the logistics cost of a single ticket; Represents the platform system usage fee; , represents the allocation weight of each cost item;

[0034] Then, the profit of participant p in the i-th transaction is calculated according to the cost formula, which is:

[0035]

[0036] in represents the total revenue in the i-th transaction, Represents the total cost of i transactions, which is used to calculate the total profit of the I transaction .

[0037] Preferably, the fraudulent transaction filtering The calculation model introduces the Mahalanobis distance model , its formula is expressed as;

[0038]

[0039] in Characteristic data representing a single transaction; Baseline characteristics of arm's length transactions; Correlation measures between features; The threshold value for fraud determination; Indicates the transaction validity flag. Output 1 indicates normal, and output 0 indicates fraud.

[0040] Preferably, in said S4, when performing verification and analysis optimization, the sensitivity of parameters is analyzed by introducing a dynamic optimization mechanism, which mainly focuses on the channel contribution adjustment factor. Parameter α in, external factor adjustment By optimizing the parameter β in the algorithm and the threshold θ in fraud detection, the clearing rule model is given the function of autonomous learning and optimization through the design of the parameter optimization model.

[0041] Preferably, for the optimization of α in the channel contribution adjustment factor, the calculation formula is:

[0042]

[0043] in represents the learning rate; Indicates the current cycle, Indicates the next cycle; is the loss gradient, which represents the influence of the current α on the error;

[0044] For the parameter β optimization in external factor adjustment, the calculation formula is:

[0045]

[0046] in represents the income loss caused by the risk in period t; K represents the risk bearing coefficient, , is the lower limit of the coefficient; The upper limit of the coefficient is updated as follows: when β rises, the risk penalty is strengthened; when β falls, the risk penalty is relaxed; when β remains unchanged, the current strategy is maintained.

[0047] The objective function used to optimize the threshold θ in fraud detection can be expressed as;

[0048]

[0049] in , indicating customer experience impairment; , indicating the risk of capital loss; Represents the area under the ROC curve, which represents the comprehensive discrimination ability of the model, where , in different business scenarios by changing The value of is used to optimize the risk of financial loss caused by the fraud risk factor.

[0050] The beneficial effects are as follows:

[0051] 1. The method for clearing ticket revenue based on multivariate heterogeneous data establishes a clearing rule model, designs an independent function model based on the contribution content of each participant, designs the contribution content as a contribution factor, and uniformly incorporates the design into the clearing rule model. In this way, under the premise of the basic sharing ratio agreed in the contract, according to the performance of each participant in the contribution content for which it is responsible, the data is normalized to form calculable specific parameterized data, and then the contribution factor of the participant is expressed by the data according to the calculation based on the data content. Then, the basic sharing ratio is floated according to the assignment, thereby mobilizing the enthusiasm of the participants and bringing better performance in the contribution content for which they are responsible. At the same time, when mistakes occur in the contribution content for which they are responsible, they will be punished and held accountable, thereby forming a positive promotion effect.

[0052] 2. This method of clearing ticket revenue based on multivariate heterogeneous data analyzes parameter sensitivity by introducing a dynamic optimization mechanism. The dynamic optimization mechanism mainly focuses on the optimization of parameter α in the channel contribution adjustment factor, parameter β in the external factor adjustment, and threshold θ in fraud detection. By designing a parameter optimization model, the parameters are optimized according to the deviation between the final distribution result and the expected distribution result caused by the impact of the parameters in each cycle, so that the clearing rule model has the function of autonomous learning and optimization, and can make the actual distribution more accurate and reasonable. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0054] Figure 1 This is a flow chart of the ticket revenue clearing method of the present invention;

[0055] Figure 2 This is a schematic diagram of multiple data sources of the present invention;

[0056] Figure 3 Schematic diagram of data preprocessing and feature normalization of the present invention;

[0057] Figure 4 This is a schematic diagram of the clearing rule modeling and calculation functional architecture of the present invention;

[0058] Figure 5 This is a flow chart for clearing execution and verification optimization of the present invention. DETAILED DESCRIPTION

[0059] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention are clearly and completely described. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. All other embodiments derived by persons of ordinary skill in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0060] In order to better understand the above technical solution, the above technical solution will be described in detail below with reference to the accompanying drawings and specific implementation methods.

[0061] The embodiment of the present invention discloses a method for clearing ticket revenue based on multivariate heterogeneous data. Figure 1-5 As shown, the following steps are included:

[0062] S1. Data collection: lock users through the ticketing system and obtain multi-dimensional data information about users;

[0063] S2. Data preprocessing and feature normalization: By designing a multivariate data pipeline, we clean the ticketing information-related data generated by users and then extract executable data features from each type of data.

[0064] S3. Establish a clearing rule model, the formula of which is:

[0065]

[0066] in Indicates the final share ratio of participant p; Indicates the basic share ratio of participant p, which is determined by the previous contractual constraints of the participants; represents the channel contribution factor; represents external influencing factors; represents attendance factor;

[0067] Then according to the share ratio of participant p , substituted into the single transaction share to calculate the total profit share obtained by the participants in the i-th transaction, and the formula can be expressed as;

[0068]

[0069] in Indicates the net profit of a single transaction; Indicates fraudulent transaction filtering;

[0070] S4. Clearing execution and verification optimization: By bringing actual ticketing revenue and related data into the model calculation, the specific value of the profit sharing of each participant is obtained. The specific data is then compared with the expected sharing ratio of the participants to verify whether it is reasonable, and the model parameters are optimized based on the analysis results.

[0071] Preferably, in the data collection process, by identifying and accessing multiple data sources, including ticketing systems and transaction systems of various distribution platforms, transaction data can be directly obtained, and then the user information in the transaction data is used to conduct big data analysis on the user to obtain user behavior data and user portrait data. At the same time, the actual actual usage data is obtained by combining on-site equipment, external environment data and text feedback data are obtained through the external Internet, and the contract terms data are manually entered, mainly for the basic share ratio of each participant. Then, by designing a cost calculation model and analyzing the cost data, we can use it to calculate .

[0072] Preferably, in step S2, the data cleaning process is performed by establishing reusable cleaning rules and pipelines, and then processing the data, including:

[0073] Structured data, handling missing values, including interpolation, deletion, and labeling; outlier detection and processing, such as abnormally high ticket prices and negative quantities; data type conversion; unique identifier generation and matching, including unifying channel IDs;

[0074] Semi-structured data: parse ticket order data, extract key fields and structure them.

[0075] Unstructured text data is cleaned, keywords are extracted, and normalized into positive or negative representations based on keywords. For example, keywords such as "difficulty getting tickets" and "poor seating" are identified and judged.

[0076] Time series data, processing timestamp formats, aligning time series, and processing out-of-order data, such as aligning user browsing log times with transaction times;

[0077] Spatial data, standardize geographic coordinates, and calculate the distance between the ticket purchase location and the venue.

[0078] Preferably, in S3, the contribution factor to the channel Normalized data processing is performed to obtain the calculation model of the channel contribution factor, and its formula is:

[0079]

[0080] in represents the quantitative value of the contribution of channel c, represents the normalized contribution ratio of channel c, the weight α=0.3 represents the control parameter of the reward and punishment range, and b=0.5 is the expected contribution ratio threshold; Represents the final channel contribution factor of the participants related to channel c.

[0081] Preferably, in S3, the external influencing factors Perform normalization processing to obtain its calculation model, and its formula is:

[0082]

[0083] in , represents the penalty coefficient for the responsible party, Represents the negative sentiment index, which belongs to the negative sentiment data in unstructured text data, including the sentiment evaluation of ticketing system, field, and environmental factors, and is normalized according to positive or negative; , represents the risk sensitivity coefficient; An exponential decay function representing the negative sentiment index.

[0084] Preferably, attendance factor The model calculation formula is:

[0085]

[0086] Where r represents the actual attendance rate; Indicates the minimum share ratio. is the threshold parameter, trigger thresholds for penalties; The threshold for full share.

[0087] In the actual allocation process, the parties affected by attendance rate are mainly the organizers or venues. When r≤60%, the organizer's share is reduced to 70%; when 60%≤r≤90%, the share is adjusted according to the linear ratio;

[0088] When r>90%, the organizer receives the full amount of its pre-allocated proportion, thereby promoting the organizer to mobilize participation in the entire event through attendance rate, thereby increasing overall ticket revenue.

[0089] Preferably, the net income from a single transaction When calculating, by introducing the cost calculation model, the formula is expressed as;

[0090]

[0091] in represents the total cost borne by participant p; Indicates the single transaction fee; Indicates the logistics cost of a single ticket; Represents the platform system usage fee; , represents the allocation weight of each cost item;

[0092] Then, the profit of participant p in the i-th transaction is calculated according to the cost formula, which is:

[0093]

[0094] in represents the total revenue in the i-th transaction, Represents the total cost of i transactions, which is used to calculate the total profit of the I transaction .

[0095] Optimally, fraudulent transaction filtering The calculation model introduces the Mahalanobis distance model , its formula is expressed as;

[0096]

[0097] in Characteristic data representing a single transaction; Baseline characteristics of arm's length transactions; Correlation measures between features; The threshold value for fraud determination; Indicates the transaction validity flag. Output 1 indicates normal, and output 0 indicates fraud.

[0098] Preferably, in S4, when performing verification and analysis optimization, the sensitivity of parameters is analyzed by introducing a dynamic optimization mechanism, which mainly focuses on the channel contribution adjustment factor. Parameter α in, external factor adjustment By optimizing the parameter β in the algorithm and the threshold θ in fraud detection, the clearing rule model is given the function of autonomous learning and optimization through the design of the parameter optimization model.

[0099] Preferably, for the optimization of α in the channel contribution adjustment factor, the calculation formula is:

[0100]

[0101] in represents the learning rate; Indicates the current cycle, Indicates the next cycle; is the loss gradient, which represents the influence of the current α on the error;

[0102] For the parameter β optimization in external factor adjustment, the calculation formula is:

[0103]

[0104] in represents the income loss caused by the risk in period t; K represents the risk bearing coefficient, , is the lower limit of the coefficient; The upper limit of the coefficient is updated as follows: when β rises, the risk penalty is strengthened; when β falls, the risk penalty is relaxed; when β remains unchanged, the current strategy is maintained.

[0105] The objective function used to optimize the threshold θ in fraud detection can be expressed as;

[0106]

[0107] in , indicating customer experience impairment; , indicating the risk of capital loss; Represents the area under the ROC curve, which represents the comprehensive discrimination ability of the model, where , in different business scenarios by changing The value of is used to optimize the risk of financial loss caused by the fraud risk factor.

[0108] Working principle: In this method of clearing ticket revenue based on multivariate heterogeneous data, by establishing a clearing rule model, an independent function model is designed based on the contribution content of each participant, and the contribution content is designed as a contribution factor, and the unified design is included in the clearing rule model. In this way, under the premise of the basic sharing ratio agreed in the contract, according to the performance of each participant in the contribution content for which it is responsible, the data is normalized to form calculable specific parameterized data, and then calculated according to the data content, so as to express the assignment of the participant's contribution factor through the data, and then float on the basic sharing ratio according to the assignment, thereby mobilizing the enthusiasm of the participants and bringing better performance in the contribution content for which they are responsible. At the same time, when mistakes occur in the contribution content for which they are responsible, they will be punished and held accountable, thereby forming a positive promotion effect.

[0109] Secondly, by introducing a dynamic optimization mechanism, the parameter sensitivity is analyzed. The dynamic optimization mechanism mainly focuses on the optimization of parameter α in the channel contribution adjustment factor, parameter β in the external factor adjustment, and threshold θ in fraud detection. By designing a parameter optimization model, the parameters are optimized according to the deviation between the final distribution result and the expected distribution result caused by the impact of the parameters in each cycle, so that the clearing rule model has the function of autonomous learning and optimization, and can make the actual distribution more accurate and reasonable.

[0110] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply the existence of any such actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article, or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or device. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, method, article, or device comprising the element.

[0111] The basic principles, main features, and advantages of the present invention are shown and described above. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions are merely illustrative of the principles of the present invention. Various changes and modifications may be made to the present invention without departing from the spirit and scope of the present invention. Such changes and modifications are intended to fall within the scope of the present invention. The scope of protection claimed in the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for clearing ticket revenue based on multivariate heterogeneous data, characterized in that: The following steps are involved: S1. Data collection: lock users through the ticketing system and obtain multi-dimensional data information about users; S2. Data preprocessing and feature normalization: By designing a multivariate data pipeline, we clean the ticketing information-related data generated by users and then extract executable data features from each type of data. S3. Establish a clearing rule model, the formula of which is: in Indicates the final share ratio of participant p; Indicates the basic share ratio of participant p, which is determined by the previous contractual constraints of the participants; represents the channel contribution factor; represents external influencing factors; represents attendance factor; Then according to the share ratio of participant p Substituting this into the single transaction share calculation, we can get the total profit share that the participants receive in the i-th transaction. The formula can be expressed as: in Indicates the net profit of a single transaction; Indicates fraudulent transaction filtering; S4. Clearing execution and verification optimization: By inputting actual ticketing revenue and related data into the model calculation, the specific profit sharing values ​​of each participant are obtained. The specific data is then compared with the expected sharing ratio of the participants to verify whether it is reasonable, and the model parameters are optimized based on the analysis results; During the data collection process, the transaction data is directly obtained by identifying and accessing multiple data sources, including the ticketing system and the transaction systems of various distribution platforms. Then, the user information in the transaction data is used to conduct big data analysis on the user to obtain user behavior data and user portrait data. At the same time, actual usage data is obtained by combining on-site equipment, external environment data and text feedback data are obtained through the external Internet, and contract terms data, including the basic share ratio of each participant, are manually entered. Then, by designing a cost calculation model and analyzing the cost data, we can use it to calculate ; In step S2, the data cleaning process is performed by establishing reusable cleaning rules and pipelines, and then processing the data, including: Structured data, handling missing values, outlier detection and processing, data type conversion, unique identifier generation and matching; Semi-structured data: parsing ticket order data, extracting key fields and structuring them; Unstructured text data, text cleaning, extracting keywords from the text and normalizing them into positive or negative values ​​according to the keywords; Time series data, processing timestamp formats, aligning time series, and processing out-of-order data; Spatial data: standardize geographic coordinates and calculate the distance between the ticket purchase location and the venue; The fraudulent transaction filtering The calculation model introduces the Mahalanobis distance model , its formula is expressed as; in Characteristic data representing a single transaction; Indicates baseline characteristics of arm's length transactions; Represents the correlation measure between features; Indicates the boundary value of fraud determination; Indicates the transaction validity flag. Output 1 indicates normal, and output 0 indicates fraud.

2. The method for clearing ticket revenue based on multivariate heterogeneous data according to claim 1, characterized in that: In S3, the channel contribution factor Normalized data processing is performed to obtain the calculation model of the channel contribution factor, and its formula is: in represents the quantitative value of the contribution of channel c, represents the normalized contribution ratio of channel c, the weight α=0.3 represents the control parameter of the reward and punishment range, and b=0.5 is the expected contribution ratio threshold; Represents the final channel contribution factor of the participants related to channel c.

3. The method for clearing ticket revenue based on multivariate heterogeneous data according to claim 2, characterized in that: In S3, external influencing factors Perform normalization processing to obtain its calculation model, and its formula is: in , represents the penalty coefficient for the responsible party, Represents the negative sentiment index, which belongs to the negative sentiment data in unstructured text data, including the sentiment evaluation of ticketing system, venue and environmental factors, and is normalized according to positive or negative; , represents the risk sensitivity coefficient; An exponential decay function representing the negative sentiment index.

4. The method for clearing ticket revenue based on multivariate heterogeneous data according to claim 1, characterized in that: The attendance factor The model calculation formula is: Where r represents the actual attendance rate; Indicates the minimum share ratio. is the threshold parameter, trigger thresholds for penalties; The threshold for full share.

5. The method for clearing ticket revenue based on multivariate heterogeneous data according to claim 1, characterized in that: Net income from a single transaction When calculating, by introducing the cost calculation model, the formula is expressed as; in represents the total cost borne by participant p; Indicates the single transaction fee; Indicates the logistics cost of a single ticket; Represents the platform system usage fee; , represents the allocation weight of each cost item; Then, the profit of participant p in the i-th transaction is calculated according to the cost formula, which is: in represents the total revenue in the i-th transaction, Represents the total cost of i transactions, which is used to calculate the total profit of the i-th transaction .

6. The method for clearing ticket revenue based on multivariate heterogeneous data according to claim 1, characterized in that: In S4, when performing verification and analysis optimization, the sensitivity of parameters is analyzed by introducing a dynamic optimization mechanism, which focuses on the channel contribution adjustment factor. Parameter α in, external factor adjustment By optimizing the parameter β in the algorithm and the threshold θ in fraud detection, the clearing rule model is given the function of autonomous learning and optimization through the design of the parameter optimization model.

7. The method for clearing ticket revenue based on multi-dimensional heterogeneous data according to claim 6, characterized in that: For the α optimization of the channel contribution adjustment factor, the calculation formula is: in represents the learning rate; Indicates the current cycle, Indicates the next cycle; is the loss gradient, which represents the influence of the current α on the error; For the parameter β optimization in external factor adjustment, the calculation formula is: in represents the income loss caused by the risk in period t; K represents the risk bearing coefficient, , is the lower limit of the coefficient; The upper limit of the coefficient is updated as follows: when β rises, the risk penalty is strengthened; when β falls, the risk penalty is relaxed; when β remains unchanged, the current strategy is maintained. The objective function used to optimize the threshold θ in fraud detection can be expressed as; in , indicating customer experience impairment; , indicating the risk of capital loss; Represents the area under the ROC curve, which represents the comprehensive discrimination ability of the model, where , in different business scenarios by changing The value of is used to optimize the risk of financial loss caused by the fraud risk factor.

Images