Tail money pricing determination method and device, electronic equipment and computer storage medium

Abstract

The invention provides a tail money pricing determination method and device, electronic equipment and a computer storage medium, and the method comprises the steps: inputting behavior time sequence data of a user for a target commodity into a fusion model after obtaining the behavior time sequence data of the user for the target commodity, and outputting to obtain a money interval of the target commodity; wherein the fusion model is composed of a bidirectional long short-term memory network and a Transform; and finally, based on the earnest money interval of the target commodity and the game information of the participants, determining the tail money price of the target commodity. Wherein the participants comprise a supplier, a platform and a user. By coordinating the interests of multiple parties, the price acceptability of the user in the tail payment stage is effectively improved, so that the selling risk is reduced, the sales volume is improved, and greater profits are obtained.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a method, device, electronic device and computer storage medium for determining final payment pricing. Background Art

[0002] In the pre-sale scenario, the sale of goods is divided into two stages: "deposit" and "final payment". The sales cycle is longer than that of general spot goods, and the inventory and pricing strategies of goods are also more flexible.

[0003] Currently, traditional pricing methods (such as linear regression and ARIMA) rely on static data and struggle to capture the temporal characteristics of user behavior and market fluctuations. This results in delayed pricing and insufficient flexibility. Users are often unacceptable to final payment prices and refuse to pay, leading to overstocking and the risk of unsold goods. Therefore, effectively mitigating sales risks while simultaneously increasing sales and achieving greater profitability is a pressing issue for merchants on e-commerce platforms. Summary of the Invention

[0004] In view of this, the present application provides a method, device, electronic device and computer storage medium for determining the final payment pricing, which effectively improves the user's price acceptance during the final payment stage, thereby reducing sales risks and increasing sales at the same time to obtain greater profits.

[0005] The first aspect of the present application provides a method for determining the final payment pricing, comprising:

[0006] Obtain time series data on user behavior towards target products;

[0007] Input the user's behavior time series data for the target product into the fusion model, and output the deposit range of the target product; wherein the fusion model is composed of a bidirectional long short-term memory network and a Transformer;

[0008] The final payment pricing of the target product is determined based on the deposit range of the target product and the game information of the participants; wherein the participants include suppliers, platforms and users.

[0009] Optionally, inputting the user's behavior time series data for the target product into a fusion model and outputting a deposit range for the target product includes:

[0010] After receiving the user's behavior time series data on the target product, the fusion model preprocesses the user's behavior time series data on the target product to obtain standardized user data;

[0011] Inputting the standardized user data into a bidirectional long short-term memory network and outputting local features;

[0012] Input the standardized user data into Transformer and output global features;

[0013] Determining a fusion feature according to the local feature, the global feature, and the trainable parameter;

[0014] A deposit range for a target product is determined based on the fused features, the learnable weight matrix, and the bias term.

[0015] Optionally, the bidirectional long short-term memory network includes a forward long short-term memory network and a backward long short-term memory network, and the inputting of the standardized user data into the bidirectional long short-term memory network and the outputting of the local features include:

[0016] The forward long short-term memory network captures the positive behavior chain from the start time to the end time;

[0017] The backward long short-term memory network captures reverse causal information from the end time to the start time;

[0018] Hidden state splicing is performed based on the forward behavior chain and reverse causal information to obtain local features.

[0019] Optionally, the participant's gaming information includes the participant's gaming function and the participant's gaming strategy variables, and determining the final payment pricing of the target product based on the deposit range of the target product and the participant's gaming information includes:

[0020] Based on the game function of the participant and the game strategy variables of the participant, the game function value of the participant is calculated;

[0021] The Shapley value of each participant is calculated based on the combination of participants and the game function value of the participants;

[0022] The Nash equilibrium is solved based on the Shapley values ​​of all the participants to determine the final payment price of the target product.

[0023] Optionally, the method for determining the final payment pricing further includes:

[0024] If the dynamic compensation conditions are met, the default probability of the final payment will be updated based on the Bayesian theorem, and the adjusted final payment pricing of the target product will be generated based on the price volatility of the current competing products.

[0025] Optionally, the method for determining the final payment pricing further includes:

[0026] If the current price volatility of the competing product exceeds the floating threshold, the participant's shared compensation amount will be determined based on the participant's Shapley value and the adjusted final payment price of the target product.

[0027] Optionally, the method for determining the final payment pricing further includes:

[0028] The user's payment status and the participant's compensation execution status are tracked in real time, and the user's payment status and the participant's compensation execution status are fed back to the fusion model.

[0029] A second aspect of the present application provides a device for determining final payment pricing, comprising:

[0030] An acquisition unit, used to acquire time series data of user behavior towards a target product;

[0031] A deposit interval determination unit, configured to input the user's time-series behavior data for a target product into a fusion model and output the deposit interval for the target product; wherein the fusion model is composed of a bidirectional long short-term memory network and a Transformer;

[0032] The final payment pricing determination unit is used to determine the final payment pricing of the target product based on the deposit range of the target product and the game information of the participants; wherein the participants include suppliers, platforms and users.

[0033] Optionally, the deposit interval determining unit includes:

[0034] A preprocessing unit, configured to preprocess the user's behavior time series data for the target product after the fusion model receives the user's behavior time series data for the target product to obtain standardized user data;

[0035] A local feature generation unit, configured to input the standardized user data into a bidirectional long short-term memory network and output local features;

[0036] A global feature generation unit, configured to input the standardized user data into a Transformer and output a global feature;

[0037] a fusion feature generating unit, configured to determine a fusion feature based on the local feature, the global feature, and a trainable parameter;

[0038] A deposit interval determination unit is used to determine the deposit interval of the target product based on the fusion feature, the learnable weight matrix and the bias term.

[0039] Optionally, the bidirectional long short-term memory network includes a forward long short-term memory network and a backward long short-term memory network, and the local feature generation unit includes:

[0040] A first capture unit is used for the forward long short-term memory network to capture the positive behavior chain from the start time to the end time;

[0041] A second capturing unit is used for the backward long short-term memory network to capture reverse causal information from the end time to the start time;

[0042] A splicing unit is used to perform hidden state splicing based on the forward behavior chain and the reverse causal information to obtain local features.

[0043] Optionally, the participant's game information includes the participant's game function and the participant's game strategy variables, and the final payment pricing determination unit includes:

[0044] A game function value calculation unit, configured to calculate the game function value of a participant based on the participant's game function and the participant's game strategy variables;

[0045] A Shapley value calculation unit, configured to calculate the Shapley value of each participant based on the combination of participants and the game function values ​​of the participants;

[0046] The final payment pricing determination subunit is used to solve the Nash equilibrium based on the Shapley values ​​of all the participants to determine the final payment pricing of the target product.

[0047] Optionally, the device for determining the final payment pricing further includes:

[0048] The dynamic adjustment unit is used to update the default probability of the final payment based on the Bayesian theorem if the dynamic compensation conditions are met, and to generate the adjusted final payment pricing of the target product based on the price volatility of the current competing products.

[0049] Optionally, the device for determining the final payment pricing further includes:

[0050] The apportioned compensation unit is used to determine the participant's apportioned compensation amount based on the participant's Shapley value and the adjusted final payment price of the target product if the price volatility of the current competing product exceeds the floating threshold.

[0051] Optionally, the device for determining the final payment pricing further includes:

[0052] The feedback unit is used to track the user's payment status and the participant's compensation execution status in real time, and feed back the user's payment status and the participant's compensation execution status to the fusion model.

[0053] A third aspect of the present application provides an electronic device, including:

[0054] one or more processors;

[0055] a storage device having one or more programs stored thereon;

[0056] When the one or more programs are executed by the one or more processors, the one or more processors implement the method for determining the final payment pricing as described in any one of the first aspects.

[0057] A fourth aspect of the present application provides a computer storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the method for determining the final payment pricing as described in any one of the first aspects is implemented.

[0058] From the above scheme, it can be seen that the present application provides a method, device, electronic device and computer storage medium for determining the final payment pricing. After obtaining the user's behavioral time series data for the target product, the user's behavioral time series data for the target product is input into the fusion model, and the output is the deposit range of the target product; wherein, the fusion model is composed of a bidirectional long short-term memory network and a Transformer; finally, based on the deposit range of the target product and the game information of the participants, the final payment pricing of the target product is determined; wherein, the participants include suppliers, platforms and users. By coordinating the interests of multiple parties, the price acceptance of users in the final payment stage can be effectively improved, thereby reducing sales risks and increasing sales at the same time, and obtaining greater profits. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] 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 merely embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying any creative work.

[0060] Figure 1 A specific flow chart of a method for determining final payment pricing provided in an embodiment of the present application;

[0061] Figure 2 A flowchart of a method for generating and determining a deposit range for a target product provided in another embodiment of the present application;

[0062] Figure 3 A flowchart of a method for generating local features provided in another embodiment of the present application;

[0063] Figure 4 A flowchart of a method for determining the final payment pricing of a target product provided in another embodiment of the present application;

[0064] Figure 5 A schematic diagram of a device for determining final payment pricing provided in another embodiment of the present application;

[0065] Figure 6A schematic diagram of an electronic device for implementing a method for determining final payment pricing provided in another embodiment of the present application. DETAILED DESCRIPTION

[0066] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0067] As used herein, the term "including" and its variations are open-ended, i.e., "including but not limited to." The term "based on" means "based, at least in part, on." The term "one embodiment" means "at least one embodiment," the term "another embodiment" means "at least one additional embodiment," and the term "some embodiments" means "at least some embodiments." Other terms are defined in the following description.

[0068] It should be noted that the information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with the relevant laws, regulations and standards of relevant countries and regions.

[0069] It should be noted that the concepts of "first" and "second" mentioned in this application are only used to distinguish different devices, modules or units, and are not used to limit the order or interdependence of the functions performed by these devices, modules or units.

[0070] It should be noted that the modifications of "one" and "multiple" mentioned in this application are illustrative rather than restrictive. Those skilled in the art should understand that unless otherwise clearly indicated in the context, they should be understood as "one or more".

[0071] The present application embodiment provides a method for determining the final payment pricing, such as Figure 1 As shown, the specific steps include:

[0072] S101: Obtain user behavior time series data for a target product.

[0073] Among them, behavioral time series data includes but is not limited to multi-dimensional time series matrices such as page dwell time, historical add-to-cart rate, competitor price comparison frequency, and collection / add-to-cart interval, which are not limited here.

[0074] For example: User A triggered 3 price comparisons and stayed on the detail page 5 times (90 seconds on average) within 24 hours.

[0075] In the specific implementation of this application, since users may accidentally touch the button, a forget gate can be used to filter out noise data, and an input gate can be used to update the weights of key behaviors. For example, an increase in price comparison frequency corresponds to an input gate activation value greater than 0.8.

[0076] S102: Input the user's behavior time series data for the target product into the fusion model, and output the deposit range of the target product.

[0077] The fusion model consists of a bidirectional long-short-term memory network and a Transformer. This model dynamically combines short-term local features of user behavior (such as click frequency) with long-term global features (such as cross-day behavior patterns). This addresses the inadequacy of traditional time series models in modeling nonlinear relationships. By using neural time series prediction, price-sensitive windows are captured, increasing the premium rate for high-demand products.

[0078] Optionally, in another embodiment of the present application, after the fusion model receives the user's behavior time series data for the target product, an implementation method of the fusion model is as follows: Figure 2 Shown, including:

[0079] S201: After receiving the time series data of user behavior for the target product, the fusion model preprocesses the time series data of user behavior for the target product to obtain standardized user data.

[0080] Optionally, in the specific implementation of this application, the pre-processing method of behavioral time series data includes but is not limited to timestamp alignment for asynchronous data streams (such as page click events and background price comparison logs), Z-Score standardization can be used to eliminate dimensional differences, and missing values ​​can be filled by interpolation (such as filling 0 for behavior fields that are not triggered by users). There is no limitation here.

[0081] S202: Input the standardized user data into a bidirectional long short-term memory network and output local features.

[0082] Optionally, in another embodiment of the present application, the bidirectional long short-term memory network includes a forward long short-term memory network and a backward long short-term memory network, and an implementation of step S202 is as follows: Figure 3 Shown, including:

[0083] S301. The forward long short-term memory network captures the positive behavior chain from the start time to the end time.

[0084] Specifically, we capture the positive behavior chain from t=1 to t=T, for example: browse → add to cart → pay.

[0085] S302. The backward long short-term memory network captures reverse causal information from the end time to the start time.

[0086] Specifically, it captures reverse causal information from t=T to t=1, such as hesitant behavior before payment.

[0087] S303: Perform hidden state splicing based on the forward behavior chain and reverse causal information to obtain local features.

[0088] For example, the local features finally obtained, such as the hidden state of user A at t=15 minutes, include the forward click pattern and the backward payment willingness (probability), which are not limited here.

[0089] In the specific implementation process of this application, the hidden state splicing method can be adopted as follows:

[0090] ;

[0091] in, Represents the local features at time step t after hidden state splicing, represents the forward hidden state, i.e. the forward behavior chain, represents reverse causal information, Represents a vector concatenation operation, which concatenates the forward hidden state and the backward hidden state into a longer vector. This concatenated vector is the hidden state of the bidirectional LSTM. express The dimension is 100. In the specific implementation process of this application, the dimension setting can be, but is not limited to, setting the number of neurons in the forward long short-term memory network to 50 and the number of neurons in the backward long short-term memory network to 50. It can also be changed according to needs. For example, if it is more necessary to capture the user's positive behavior chain, then the number of neurons in the forward long short-term memory network can be set to 80 and the number of neurons in the backward long short-term memory network can be set to 20. There is no limitation here.

[0092] It should be noted that the above only takes 100 dimensions as an example. In the specific implementation process of this application, other dimensions can of course be selected, and this is not limited here.

[0093] S203: Input the standardized user data into the Transformer and output the global features.

[0094] In the specific implementation process of this application, Transformer uses the multi-head self-attention mechanism to calculate the global eigenvalue:

[0095] ;

[0096] Among them, Q represents the query matrix, which represents the information that currently needs attention; K represents the key matrix, which is used to match the query matrix to determine the degree of attention; V represents the value matrix, which contains the actual information associated with the key and will be weighted and summed according to the attention weight; dk represents the dimension of the key vector, which is used to scale the dot product result to prevent it from being too large and causing the gradient vanishing problem of the softmax function.

[0097] Specifically, the above expression represents the global features obtained by calculating the dot product of the transpose of the query matrix Q and the key matrix K, dividing it by the square root of the dimension dk of the key vector, and then normalizing it through the softmax function. By calculating the similarity between the query and the key and assigning attention weights accordingly, the model can process the input information more efficiently, thereby improving the overall performance.

[0098] It is understandable that since local features and global features need to be fused later, the dimensions of local features and global features must be consistent.

[0099] It should be noted that, in actual application, step S203 and step S202 may be performed sequentially or simultaneously, and the order may be S202 first and then S203, or S203 first and then S202. Figure 2 It is just an example of first performing S202 and then performing S203, and is not limited here.

[0100] S204: Determine fusion features based on local features, global features, and trainable parameters.

[0101] The trainable parameter α may range from, but is not limited to, 0.3 to 0.7, and is not limited here.

[0102] In the specific implementation process of this application, the local features and global features are weighted by trainable parameters to finally obtain the fusion feature implementation method, which can be calculated using the following formula:

[0103] Ffusion=α×FLSTM+(1−α)×FTransformer;

[0104] Among them, FLSTM represents local features, FTransformer represents global features, and Ffusion represents fusion features.

[0105] It should be noted that α can also be dynamically optimized through back propagation to achieve adaptive allocation of feature contributions.

[0106] For example, when user behavior fluctuates greatly (e.g., the standard deviation of price comparison frequency is > 1.5), the model automatically reduces α to 0.4 and enhances the global feature weight.

[0107] S205: Determine the deposit range for the target product based on the fusion features, the learnable weight matrix, and the bias term.

[0108] In the specific implementation process of this application, the deposit range of the target product is determined based on the fusion features, the learnable weight matrix, and the bias term. This can be implemented in the fully connected layer. The specific implementation method can be as follows:

[0109] [Pmin, Pmax]=W×LeakyReLU(Ffusion)+b;

[0110] Here, [Pmin, Pmax] represents the deposit range, with Pmin representing the minimum and Pmax the maximum. W is a learnable weight matrix, and the activation function uses LeakyReLU (with a negative slope of 0.01) to prevent negative gradient clipping. The high-dimensional fused features are mapped to the price range space, with different rows corresponding to the linear transformation parameters for Pmin and Pmax. b is the initial bias term, such as [￥250, ￥500], corresponding to the baseline values ​​for the upper and lower bounds of the range, respectively.

[0111] S103: Determine the final payment price of the target product based on the deposit range of the target product and the game information of the participants.

[0112] Participants include suppliers, platforms and users.

[0113] Optionally, in another embodiment of the present application, the game information of the participant includes the game function of the participant and the game strategy variable of the participant. An implementation of step S103 is as follows: Figure 4 Shown, including:

[0114] S401. Calculate the participant's game function value based on the participant's game function and the participant's game strategy variables.

[0115] It is understandable that suppliers and platforms usually aim to maximize profits, while users (consumers) aim to maximize utility. In the actual application of this application, the platform's game function can be:

[0116] U p =μ×(P deposit + P balance)×D(P) − compensation cost;

[0117] U p is the platform profit, μ is the profit rate, P deposit is the deposit at the time of transaction, P balance is the balance at the time of transaction, and D(P) is the demand function.

[0118] In the actual application process of this application, D(P) can be calculated using the following function:

[0119] ;

[0120] in, is a constant representing the maximum possible error probability, is the base of the natural logarithm, and λ is a parameter that can be, but is not limited to, 0.05. It refers to the pricing of goods. The price is the user's psychological expectation, which can be obtained by analyzing user behavior data.

[0121] The user's game function can be:

[0122] ;

[0123] in, For user utility, is the behavioral data conversion coefficient, The user's sensitivity to price.

[0124] In the specific implementation process of this application, the user's game strategy variable is , the supplier's game strategy variable is the cost threshold, and the platform's game strategy variable is the profit margin. For example, the supplier's cost threshold C min C min ≥￥800, user's ϵ∈[0.5, 1.2], platform's μ∈[0.15, 0.35].

[0125] S402: Calculate the Shapley value of each participant based on the combination of participants and the game function values ​​of the participants.

[0126] In the specific implementation process of this application, the following formula can be used to calculate the Shapley value of the participant:

[0127] ;

[0128] in, Indicates participants The Shapley value of N represents the set of all participants (such as users and platforms; suppliers, platforms and users; suppliers and platforms, etc.), , represents the total number of participants, S does not include participants Alliance, is the total number of participants - 1; is the weight factor, reflecting the probability distribution of the set size; is the total revenue of the S alliance, assuming that N includes suppliers, platforms and users, and participants For users, S Alliance is for suppliers, platforms and the combination of suppliers and platforms. Therefore, The benefits of suppliers, platforms and suppliers and platforms are determined based on the game function value. , traverse the S alliance and calculate its marginal contribution , and according to , calculate the weight, for example: when n=3, When it is 1, the weight It is 1 / 6.

[0129] S403. Solve the Nash equilibrium based on the Shapley values ​​of all participants to determine the final payment price of the target product.

[0130] In the specific implementation process of this application, if the participants are the platform and the user, the following iterative formula can be used to calculate the final payment pricing of the target product:

[0131] ;

[0132] in, exist The final payment pricing at the time of the iteration, represents the platform’s profit when the product is priced at P, represents the user utility when the product is priced at P, represents the Shapley value of the platform, Represents the user's Shapley value.

[0133] Optionally, in another embodiment of the present application, an implementation of the method for determining the final payment pricing further includes:

[0134] If the dynamic compensation conditions are met, the default probability of the final payment will be updated based on the Bayesian theorem, and the adjusted final payment pricing of the target product will be generated based on the price volatility of the current competing products.

[0135] It can be understood that the probability of default + the probability of payment = 1, where the probability of payment is the willingness to pay (probability) in the local features obtained by the output of the bidirectional long short-term memory network.

[0136] Among them, dynamic compensation conditions include but are not limited to the probability of default of the final payment exceeding the default threshold, the price volatility of the competing product exceeding the volatility threshold, and periodic triggering (such as a full parameter update every 24 hours to ensure adaptation to market dynamics), etc., which are not limited here.

[0137] In the specific implementation of this application, the price volatility of competing products can be calculated as follows:

[0138] ;

[0139] in, Indicates the price volatility of competing products, The latest price of competing products. This is the price before the competitor's latest price is updated.

[0140] Optionally, in the specific implementation process of this application, based on the price volatility of the current competing products, an implementation method for generating the adjusted final payment pricing of the target product can be obtained by the following calculation: :

[0141] =min(0.5+0.1×(ΔPcompetitor−5%), 0.8);

[0142] If ΔP competitor is 8%, then is 0.74, using the new Calculate the new , and use the new Calculate the adjusted final payment price of the target product.

[0143] Optionally, in another embodiment of the present application, an implementation of the method for determining the final payment pricing further includes:

[0144] If the current price volatility of the competing product exceeds the floating threshold, the participant's shared compensation amount will be determined based on the participant's Shapley value and the adjusted final payment price of the target product.

[0145] Continuing with the above example, if the target product's final price is 799, the adjusted target product's final price is 779, the platform's Shapley value is 0.45, the supplier's Shapley value is 0.35, and the user's Shapley value is 0.2, then the total compensation is 799-779=20, the platform's apportioned compensation amount is 0.45×20=9, the supplier's apportioned compensation amount is 0.35×20=7, and the user's apportioned compensation amount (implicitly borne, coupons can be issued to users for next consumption, not limited here) is 0.2×20=4.

[0146] This application converts potential lost customers through a difference compensation strategy, thereby reducing the platform's bad debt loss rate.

[0147] Optionally, in another embodiment of the present application, an implementation of the method for determining the final payment pricing further includes:

[0148] Track the user's payment status and the participants' compensation execution in real time, and feed the user's payment status and the participants' compensation execution back to the fusion model.

[0149] Specifically, the fusion model calculates the cross-entropy loss function based on the user's payment status and the participants' compensation execution, and uses a preset learning rate for model optimization training.

[0150] In the specific implementation process of this application, the payment method of the balance can be but is not limited to realizing the balance payment through smart contract, for example: automatically initiating the balance payment within 72 hours after the arrival of the goods, etc., which is not limited here.

[0151] From the above scheme, it can be seen that this application provides a method for determining the final payment pricing. After obtaining the user's behavioral time series data for the target product, the user's behavioral time series data for the target product is input into the fusion model, and the output is the deposit range of the target product; wherein, the fusion model is composed of a bidirectional long short-term memory network and a Transformer; finally, based on the deposit range of the target product and the game information of the participants, the final payment pricing of the target product is determined; wherein, the participants include suppliers, platforms and users. By coordinating the interests of multiple parties, the price acceptance of users in the final payment stage can be effectively improved, thereby reducing sales risks and increasing sales at the same time, and obtaining greater profits.

[0152] Another embodiment of the present application provides a device for determining the final payment pricing, such as Figure 5 As shown, specifically including:

[0153] The acquisition unit 501 is used to acquire the time series data of the user's behavior towards the target product.

[0154] The deposit interval determination unit 502 is used to input the user's behavior time series data for the target product into the fusion model and output the deposit interval of the target product.

[0155] Among them, the fusion model consists of a bidirectional long short-term memory network and a Transformer.

[0156] Optionally, in another embodiment of the present application, an implementation of the deposit interval determining unit 502 includes:

[0157] The preprocessing unit is used to preprocess the user's behavior time series data for the target product after the fusion model receives the user's behavior time series data for the target product to obtain standardized user data.

[0158] The local feature generation unit is used to input standardized user data into the bidirectional long short-term memory network and output local features.

[0159] The global feature generation unit is used to input standardized user data into the Transformer and output global features.

[0160] The fusion feature generation unit is used to determine the fusion feature according to the local features, global features and trainable parameters.

[0161] The deposit range determination unit is used to determine the deposit range of the target product based on the fusion features, the learnable weight matrix, and the bias term.

[0162] For the specific working process of the units disclosed in the above embodiments of the present application, please refer to the corresponding method embodiments, such as Figure 2 As shown, no further details are given here.

[0163] Optionally, in another embodiment of the present application, the bidirectional long short-term memory network includes a forward long short-term memory network and a backward long short-term memory network, and an implementation of the local feature generation unit includes:

[0164] The first capture unit is used to capture the positive behavior chain from the start time to the end time of the forward long short-term memory network.

[0165] The second capture unit is used to capture reverse causal information from the end time to the start time in the backward long short-term memory network.

[0166] The splicing unit is used to perform hidden state splicing based on the forward behavior chain and reverse causal information to obtain local features.

[0167] For the specific working process of the units disclosed in the above embodiments of the present application, please refer to the corresponding method embodiments, such as Figure 3 As shown, no further details are given here.

[0168] The balance price determination unit 503 is configured to determine the balance price of the target product based on the deposit range of the target product and the game information of the participants.

[0169] Participants include suppliers, platforms and users.

[0170] For the specific working process of the units disclosed in the above embodiments of the present application, please refer to the corresponding method embodiments, such as Figure 1 As shown, no further details are given here.

[0171] Optionally, in another embodiment of the present application, the participant's game information includes the participant's game function and the participant's game strategy variables. An implementation of the final payment pricing determination unit 503 includes:

[0172] The game function value calculation unit is used to calculate the game function value of the participant based on the game function of the participant and the game strategy variable of the participant.

[0173] The Shapley value calculation unit is used to calculate the Shapley value of each participant according to the combination of participants and the game function value of the participants.

[0174] The final payment pricing determination subunit is used to solve the Nash equilibrium based on the Shapley values ​​of all participants and determine the final payment pricing of the target product.

[0175] For the specific working process of the units disclosed in the above embodiments of the present application, please refer to the corresponding method embodiments, such as Figure 4 As shown, no further details are given here.

[0176] Optionally, in another embodiment of the present application, an implementation of the device for determining the final payment pricing further includes:

[0177] The dynamic adjustment unit is used to update the default probability of the final payment based on the Bayesian theorem if the dynamic compensation conditions are met, and to generate the adjusted final payment pricing of the target product based on the price volatility of the current competing products.

[0178] The specific working process of the units disclosed in the above embodiments of this application can be found in the corresponding method embodiments and will not be repeated here.

[0179] Optionally, in another embodiment of the present application, an implementation of the device for determining the final payment pricing further includes:

[0180] The apportioned compensation unit is used to determine the participant's apportioned compensation amount based on the participant's Shapley value and the adjusted final payment price of the target product if the price volatility of the current competing product exceeds the floating threshold.

[0181] The specific working process of the units disclosed in the above embodiments of this application can be found in the corresponding method embodiments and will not be repeated here.

[0182] Optionally, in another embodiment of the present application, an implementation of the device for determining the final payment pricing further includes:

[0183] The feedback unit is used to track the user's payment status and the participant's compensation execution status in real time, and feed back the user's payment status and the participant's compensation execution status to the fusion model.

[0184] The specific working process of the units disclosed in the above embodiments of this application can be found in the corresponding method embodiments and will not be repeated here.

[0185] As can be seen from the above scheme, this application provides a device for determining the final payment pricing. After the acquisition unit 501 obtains the user's behavioral time series data for the target product, the deposit interval determination unit 502 inputs the user's behavioral time series data for the target product into the fusion model, and outputs the deposit interval of the target product; wherein, the fusion model is composed of a bidirectional long short-term memory network and a Transformer; finally, the final payment pricing determination unit 503 determines the final payment pricing of the target product based on the deposit interval of the target product and the game information of the participants; wherein, the participants include suppliers, platforms and users. By coordinating the interests of multiple parties, the price acceptance of users in the final payment stage can be effectively improved, thereby reducing sales risks and increasing sales at the same time, and obtaining greater profits.

[0186] The functions described above herein may be performed, at least in part, by one or more hardware logic components. For example, and without limitation, exemplary types of hardware logic components that may be used include: field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), complex programmable logic devices (CPLDs), and the like.

[0187] Another embodiment of the present application provides an electronic device, such as Figure 6 Shown, including:

[0188] One or more processors 601 .

[0189] The storage device 602 stores one or more programs.

[0190] When the one or more programs are executed by the one or more processors 601 , the one or more processors 601 implement the method for determining the final payment pricing as described in the above embodiment.

[0191] Another embodiment of the present application provides a computer storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the method for determining the final payment pricing as described in the above embodiment is implemented.

[0192] In the context of this application, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or apparatus. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing. More specific examples of machine-readable storage media may include an electrical connection based on one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), optical fibers, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0193] It should be noted that the computer-readable medium referred to in this application may be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. Computer-readable storage media may include, but are not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or components, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to, an electrical connection having one or more conductors, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In this application, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, device, or component. In this application, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. This propagated data signal may take a variety of forms, including, but not limited to, electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium that can transmit, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. Program code embodied on a computer-readable medium may be transmitted using any suitable medium, including but not limited to wire, optical cable, RF (radio frequency), or any suitable combination thereof.

[0194] The computer-readable medium may be included in the electronic device, or may exist independently without being incorporated into the electronic device.

[0195] Another embodiment of the present application provides a computer program product, which, when executed, is used to perform the above-mentioned method for determining the final payment pricing.

[0196] In particular, according to an embodiment of the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present application includes a computer program product, which includes a computer program carried on a non-transitory computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network via a communication device, or installed from a storage device, or installed from a ROM. When the computer program is executed by a processing device, the above-mentioned functions defined in the method of the embodiment of the present application are performed.

[0197] Although the subject matter has been described in terms of structural features and / or method logic actions, it should be understood that the subject matter defined in this application is not necessarily limited to the specific features or actions described above. On the contrary, the specific features and actions described above are merely example forms of implementing this application.

[0198] Although several specific implementation details are included in the above discussion, these should not be construed as limiting the scope of this application. Certain features described in the context of separate embodiments may also be implemented in combination in a single embodiment. Conversely, various features described in the context of a single embodiment may also be implemented in multiple embodiments individually or in any suitable sub-combination.

[0199] The above description is merely a preferred embodiment of the present application and an illustration of the technical principles employed. Those skilled in the art should understand that the scope of application of this application is not limited to technical solutions formed by a specific combination of the aforementioned technical features. It also encompasses other technical solutions formed by any combination of the aforementioned technical features or their equivalents, without departing from the aforementioned application concept. For example, a technical solution formed by replacing the aforementioned features with (but not limited to) technical features with similar functions claimed in this application.

Claims

1. A method for determining the final payment pricing, characterized in that: include: Obtain time series data on user behavior towards target products; Input the user's behavior time series data for the target product into the fusion model, and output the deposit range of the target product; wherein the fusion model is composed of a bidirectional long short-term memory network and a Transformer; The final payment pricing of the target product is determined based on the deposit range of the target product and the game information of the participants; wherein the participants include suppliers, platforms and users.

2. The method for determining the final payment pricing according to claim 1, characterized in that: The step of inputting the user's behavior time series data for the target product into the fusion model and outputting the deposit range for the target product includes: After receiving the user's behavior time series data on the target product, the fusion model preprocesses the user's behavior time series data on the target product to obtain standardized user data; Inputting the standardized user data into a bidirectional long short-term memory network and outputting local features; Input the standardized user data into Transformer and output global features; Determining a fusion feature according to the local feature, the global feature, and the trainable parameter; A deposit range for a target product is determined based on the fused features, the learnable weight matrix, and the bias term.

3. The method for determining the final payment pricing according to claim 2, characterized in that: The bidirectional long short-term memory network includes a forward long short-term memory network and a backward long short-term memory network. The standardized user data is input into the bidirectional long short-term memory network, and the local features are output, including: The forward long short-term memory network captures the positive behavior chain from the start time to the end time; The backward long short-term memory network captures reverse causal information from the end time to the start time; Hidden state splicing is performed based on the forward behavior chain and reverse causal information to obtain local features.

4. The method for determining the final payment pricing according to claim 1, wherein: The participant's game information includes the participant's game function and the participant's game strategy variable. The determining of the final payment pricing of the target product based on the deposit range of the target product and the participant's game information includes: Based on the game function of the participant and the game strategy variables of the participant, the game function value of the participant is calculated; The Shapley value of each participant is calculated based on the combination of participants and the game function value of the participants; The Nash equilibrium is solved based on the Shapley values ​​of all the participants to determine the final payment price of the target product.

5. The method for determining the final payment pricing according to claim 4, characterized in that: Also includes: If the dynamic compensation conditions are met, the default probability of the final payment will be updated based on the Bayesian theorem, and the adjusted final payment pricing of the target product will be generated based on the price volatility of the current competing products.

6. The method for determining the final payment pricing according to claim 5, characterized in that: Also includes: If the current price volatility of the competing product exceeds the floating threshold, the participant's shared compensation amount will be determined based on the participant's Shapley value and the adjusted final payment price of the target product.

7. The method for determining the final payment pricing according to claim 6, characterized in that: Also includes: The user's payment status and the participant's compensation execution status are tracked in real time, and the user's payment status and the participant's compensation execution status are fed back to the fusion model.

8. A device for determining the final payment price, characterized in that: include: An acquisition unit, used to acquire time series data of user behavior towards a target product; A deposit interval determination unit, configured to input the user's time-series behavior data for a target product into a fusion model and output the deposit interval for the target product; wherein the fusion model is composed of a bidirectional long short-term memory network and a Transformer; The final payment pricing determination unit is used to determine the final payment pricing of the target product based on the deposit range of the target product and the game information of the participants; wherein the participants include suppliers, platforms and users.

9. An electronic device, characterized in that: include: one or more processors; a storage device having one or more programs stored thereon; When the one or more programs are executed by the one or more processors, the one or more processors implement the method for determining the final payment pricing as described in any one of claims 1 to 7.

10. A computer storage medium, characterized in that A computer program is stored thereon, wherein when the computer program is executed by a processor, the method for determining the final payment pricing as described in any one of claims 1 to 7 is implemented.

Images