A method and system for generating intelligent commodity pricing strategies that integrates multi-factor analysis

Abstract

This invention relates to the field of intelligent product pricing technology, and discloses a method and system for generating intelligent product pricing strategies by integrating multi-factor analysis. The method involves: calculating a weight set and multiple first pricing parameters based on the historical pricing records of each tenant in the rights and benefits platform; performing a weighted operation on the multiple first pricing parameters based on the weight set to obtain the tenant's second pricing parameters; multiplying the second pricing parameters by the category benchmark price to generate M first candidate prices; verifying each of the M first candidate prices, selecting N second candidate prices that satisfy all constraints, calculating the comprehensive score of each second candidate price, and selecting the second candidate price with the highest comprehensive score as the target price and storing it in the tenant pricing matrix. This invention solves the technical problem in existing technologies where all tenants share the same pricing weights, leading to an inability to adapt to differentiated business needs, by integrating multi-factor analysis, and achieves intelligent generation of product pricing strategies.

Description

Technical Field

[0001] This invention relates to the field of intelligent product pricing technology, and in particular to an intelligent product pricing strategy generation method and system that integrates multi-factor analysis. Background Technology

[0002] In the pricing process of intelligent rights and benefits platforms, traditional single-dimensional pricing methods such as cost-plus pricing, which calculate prices based on only a single factor—product cost—are ill-suited to the complex environment where multiple tenants on the platform have significantly different sensitivities to pricing factors. Each tenant has a different business model and varying degrees of emphasis on cost control, profit targets, and user value. Using a uniform global pricing rule cannot meet the differentiated needs of tenants. Furthermore, existing technologies that build independent pricing systems for each product category result in high maintenance costs and the inability to reuse pricing logic. Summary of the Invention

[0003] The main objective of this invention is to provide an intelligent product pricing strategy generation method and system that integrates multi-factor analysis. This invention solves the technical problem in the prior art where all tenants share the same pricing weight, which makes it impossible to adapt to differentiated business needs, by integrating multi-factor analysis, and realizes the intelligent generation of product pricing strategies.

[0004] To achieve the above objectives, this invention provides an intelligent product pricing strategy generation method integrating multi-factor analysis, comprising the following steps:

[0005] Calculate a weight set and multiple primary pricing parameters based on the historical pricing records of each tenant in the rights and interests platform;

[0006] Based on the weight set, a weighted operation is performed on the plurality of first pricing parameters to obtain the tenant's second pricing parameters;

[0007] The category benchmark price is calculated based on the product category label, and the second pricing parameter is multiplied by the category benchmark price to generate M first candidate prices;

[0008] Each of the M first candidate prices is verified, and N second candidate prices that satisfy all constraints are selected. The comprehensive score of each second candidate price is calculated, and the second candidate price with the highest comprehensive score is selected as the target price and stored in the tenant pricing matrix.

[0009] Optionally, in a first implementation of the first aspect of the present invention, the step of calculating a weight set and multiple first pricing parameters based on the historical pricing records of each tenant in the rights platform includes:

[0010] Obtain the historical pricing records for each tenant on the rights and interests platform, including the face value of the product, the actual selling price, the supplier's cost, and the user's discount rate;

[0011] Calculate the ratio of the supplier's cost to the actual selling price to obtain the cost percentage, and calculate the difference between the actual selling price and the supplier's cost, then divide it by the actual selling price to obtain the profit percentage;

[0012] The cost percentage is averaged to obtain the cost weight, and the profit percentage is averaged to obtain the profit weight. Default values ​​are set for the supply chain weight, user weight, scenario weight, and time weight based on the tenant's business type.

[0013] Construct a weight set using the cost weight, the profit weight, the supply chain weight, the user weight, the scenario weight, and the time weight;

[0014] Calculate the user price sensitivity coefficient, product value index, scenario conversion rate coefficient, supply chain risk coefficient, operation strategy weight, and holiday bonus coefficient to generate multiple primary pricing parameters.

[0015] Optionally, in a second implementation of the first aspect of the present invention, the step of calculating user price sensitivity coefficient, product value index, scenario conversion rate coefficient, supply chain risk coefficient, operational strategy weight, and holiday bonus coefficient to generate multiple first pricing parameters includes:

[0016] The price sensitivity coefficient of users is obtained by calculating the price variance and order frequency of users' historical orders from the user behavior record table. The commodity value index is calculated by extracting supplier quotations, write-off rates, inventory quantities and remaining validity periods from the equity SKU management table. The scenario conversion rate coefficient is calculated by statistically analyzing the historical conversion rates of cash withdrawal scenarios, repayment scenarios and loan rejection page scenarios from the order transaction table. The supply chain risk coefficient is calculated by extracting quotation volatility and supply stability from the supply chain management table. The operation strategy weight is read from the tenant configuration table to obtain the holiday bonus coefficient.

[0017] The user price sensitivity coefficient, the product value index, the scenario conversion rate coefficient, the supply chain risk coefficient, the operation strategy weight, and the holiday bonus coefficient are used as multiple first pricing parameters.

[0018] Optionally, in a third implementation of the first aspect of the present invention, the step of performing a weighted operation on the plurality of first pricing parameters according to the weight set to obtain the tenant's second pricing parameters includes:

[0019] The tenant's weight set is read, and the cost weight in the weight set is assigned to the commodity value index in the plurality of first pricing parameters. The user weight is assigned to the user price sensitivity coefficient, the scenario weight is assigned to the scenario conversion rate coefficient, the supply chain weight is assigned to the supply chain risk coefficient, the profit weight is assigned to the operation strategy weight, and the time weight is assigned to the holiday bonus coefficient. A weighted calculation is performed to obtain the tenant's second pricing parameters.

[0020] Optionally, in a fourth implementation of the first aspect of the present invention, the step of calculating a category benchmark price based on the product category label and multiplying the second pricing parameter by the category benchmark price to generate M first candidate prices includes:

[0021] Query the category tags of the target product from the main product table, determine whether the product is a rights, insurance, mall, short drama or notary product based on the category tags, and construct a routing relationship table between the category tags and the corresponding mapping functions;

[0022] The mapping function corresponding to the category tag is selected according to the routing relationship table, and the category benchmark price of the product is calculated based on the category pricing parameters corresponding to the mapping function;

[0023] The second pricing parameter is multiplied by the category benchmark price, and M first candidate prices are generated based on the category benchmark price by a preset adjustment range.

[0024] Optionally, in a fifth implementation of the first aspect of the present invention, the step of selecting the mapping function corresponding to the category tag according to the routing relationship table, and calculating the category benchmark price of the product based on the category pricing parameters corresponding to the mapping function, includes:

[0025] The mapping function corresponding to the category tag is queried from the routing table. For the rights category, it is mapped to the point multiplication function; for the insurance category, it is mapped to the coverage multiplication function; for the mall category, it is mapped to the cost division function; for the short drama category, it is mapped to the episode multiplication function; and for the notary category, it is mapped to the amount multiplication function.

[0026] The mapping function is called to substitute the corresponding category pricing parameters and perform the calculation to obtain the category benchmark price of the product.

[0027] Optionally, in a sixth implementation of the first aspect of the present invention, the step of multiplying the second pricing parameter by the category benchmark price and generating M first candidate prices based on the category benchmark price by a preset adjustment range includes:

[0028] The base price is obtained by multiplying the second pricing parameter by the benchmark price of the product category.

[0029] The difference between the scene conversion rate coefficient and the platform average conversion rate is calculated, divided by the platform average conversion rate, multiplied by the preset scene sensitivity parameter, and then one is added to obtain the scene adjustment coefficient.

[0030] Multiply the base value by the scenario adjustment coefficient to obtain the scenario adjustment price;

[0031] Extract the price change rate and sales change rate of the tenant under the product from the order history database. Divide the sales change rate by the price change rate and take the average value to obtain the price elasticity coefficient. Multiply the scenario adjustment price by the correction factor corresponding to the price elasticity coefficient to obtain the elasticity correction price.

[0032] Based on the elastically adjusted price, M first candidate prices are calculated according to a preset adjustment range.

[0033] Optionally, in the seventh implementation of the first aspect of the present invention, the step of verifying the M first candidate prices, filtering out N second candidate prices that satisfy all constraints, calculating the comprehensive score of each second candidate price, and selecting the second candidate price with the highest comprehensive score as the target price and storing it in the tenant pricing matrix includes:

[0034] For the M first candidate prices, verify in sequence whether they are greater than or equal to the lower bound of the cost difference between the supplier's quotation and the channel commission ratio, whether the product of the candidate price and the expected conversion rate and the predicted exposure is greater than or equal to the tenant's profit target, whether the expected sales volume is less than or equal to the inventory quantity, and whether the absolute value of the difference between the candidate price and the competitor's price divided by the competitor's price is less than or equal to the preset deviation threshold, and select N second candidate prices that simultaneously satisfy all constraints.

[0035] For the N second candidate prices, calculate the expected revenue, predicted refund rate, and user satisfaction. Then, perform a weighted summation of the expected revenue, the predicted refund rate, and the user satisfaction to obtain a comprehensive score for each second candidate price.

[0036] Iterate through the N second candidate prices and select the one with the highest comprehensive score as the target price. Store the target price in the tenant pricing matrix using the tenant identifier, product category identifier, and scenario identifier as index keys.

[0037] Optionally, in the eighth implementation of the first aspect of the present invention, the step of sequentially verifying whether the M first candidate prices are greater than or equal to the lower bound of the cost difference between the supplier's quotation and the channel commission ratio, whether the product of the candidate price and the expected conversion rate and the predicted exposure is greater than or equal to the tenant's profit target, whether the expected sales volume is less than or equal to the inventory quantity, and whether the absolute value of the difference between the candidate price and the competitor's price divided by the competitor's price is less than or equal to a preset deviation threshold, and selecting N second candidate prices that simultaneously satisfy all constraints, includes:

[0038] Read supplier quotes from the equity SKU management table and channel commission rates from the tenant configuration table. Calculate the difference between supplier quotes and channel commission rates to obtain the lower limit of costs. Read tenant profit targets from the tenant configuration table and inventory quantities from the inventory management table. Obtain competitor prices and preset deviation thresholds from the competitor monitoring system.

[0039] For the M first candidate prices, sequentially verify whether they are greater than or equal to the lower bound of the cost difference between the supplier's quotation and the channel commission ratio, whether the product of the candidate price and the expected conversion rate and the predicted exposure is greater than or equal to the tenant's profit target, whether the expected sales volume is less than or equal to the inventory quantity, and whether the absolute value of the difference between the candidate price and the competitor's price divided by the competitor's price is less than or equal to the preset deviation threshold, and select N second candidate prices that simultaneously satisfy all constraints.

[0040] This invention also provides an intelligent product pricing strategy generation system that integrates multi-factor analysis, comprising:

[0041] The pricing module is used to calculate a weight set and multiple primary pricing parameters based on the historical pricing records of each tenant in the rights platform;

[0042] The calculation module is used to perform a weighted calculation on the plurality of first pricing parameters according to the weight set to obtain the second pricing parameters of the tenant;

[0043] The benchmark price calculation module is used to calculate the category benchmark price based on the product category label, and multiply the second pricing parameter by the category benchmark price to generate M first candidate prices;

[0044] The filtering module is used to verify the M first candidate prices respectively, filter the N second candidate prices that satisfy all constraints, calculate the comprehensive score of each second candidate price, and select the second candidate price with the highest comprehensive score as the target price and store it in the tenant pricing matrix.

[0045] In summary, this invention constructs a multi-tenant isolated pricing architecture based on Redis sharding, allocating independent storage space to each tenant and calculating a tenant-specific six-dimensional factor weight set based on historical pricing records. This allows different tenants to obtain differentiated fusion pricing parameters for the same user and the same product according to their own business strategies, solving the technical problem in existing technologies where all tenants share the same pricing weights, making it impossible to adapt to differentiated business needs. By extracting pricing parameters from user behavior, product attributes, scenario conversion, supply chain status, tenant strategies, and time periods and performing weighted calculations, comprehensive fusion analysis of multi-dimensional factors is achieved. By designing a category adaptation layer, the heterogeneous pricing rules of five product categories—rights, insurance, e-commerce, short dramas, and notary services—are converted into a unified price space. Point multiplication functions, coverage multiplication functions, cost division functions, episode multiplication functions, and amount multiplication functions are used to achieve the reuse and extension of cross-category pricing logic, solving the technical problem in existing technologies of building independent pricing systems for each category. The pricing mechanism dynamically adjusts the price by multiplying the integrated pricing parameters by the category benchmark price and then combining the scenario adjustment coefficient and price elasticity coefficient. It also verifies the candidate price against four types of constraints: lower bound of execution cost, profit target, inventory quantity, and competitive deviation. After screening feasible prices, it calculates the comprehensive score of expected revenue, predicted refund rate, and user satisfaction to select the optimal pricing. Attached Figure Description

[0046] Figure 1 This is a schematic diagram of the steps in an embodiment of the present invention for generating an intelligent commodity pricing strategy that integrates multi-factor analysis;

[0047] Figure 2 This is a block diagram of an intelligent commodity pricing strategy generation system that integrates multi-factor analysis in one embodiment of the present invention.

[0048] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0049] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0050] Reference Figure 1 This embodiment provides an intelligent product pricing strategy generation method that integrates multi-factor analysis, including the following steps:

[0051] S1, calculates a weight set and multiple first pricing parameters based on the historical pricing records of each tenant in the rights platform;

[0052] The process involves retrieving pricing records for each tenant from the platform database over a past period (e.g., the last 90 days), including fields such as product face value, actual selling price, supplier costs, and user discount rates. The supplier cost in each historical record is divided by the actual selling price to obtain the cost percentage for that record. Simultaneously, the profit percentage is calculated by subtracting the supplier cost from the actual selling price and then dividing again. The average cost percentage across all historical records is used to obtain the cost weight, and similarly, the average profit percentage is used as the profit weight. After historical fitting of the cost and profit dimensions, a rule-based mapping is applied based on the tenant's business type. For example, if a tenant focuses on new user acquisition, the user weight is set to 0.25; if they prioritize promotional conversion, the scenario weight is preset to 0.2; if their supply chain exhibits significant fluctuations, the supply chain weight is set to 0.12; and the time weight is assigned within a range of 0.05 to 0.15. The cost weight, profit weight, supply chain weight, user weight, scenario weight, and time weight are then encapsulated into a unified weight set. Simultaneously, to construct the first pricing parameter set, order price sequences of target users are extracted from the user behavior database. A user price sensitivity coefficient is calculated based on the ratio of standard deviation to mean, and a lifetime value indicator is generated by weighting the consumption amounts of multiple categories such as video memberships, food delivery coupons, and phone bill coupons. A decay model is constructed based on supplier quotes, redemption rates, inventory quantity, and expiration dates to calculate the product value index. A scenario conversion rate coefficient is generated by comparing the CVR of each business scenario with the platform's average CVR, used to characterize the scenario's elasticity to price adjustments. A supply chain risk coefficient is derived from quote volatility and fulfillment stability to reflect cost-side uncertainty. Operational strategy weights are set according to the "customer acquisition - user activation - revenue generation" triple ratio in the tenant configuration table. The holiday bonus coefficient in the time factor needs to be obtained from a table, such as 0.5 for Spring Festival, 0.2 for short holidays, and 0 for ordinary days, and the time weight is adjusted based on holiday sales forecasts. These multiple dimensions of factors together constitute the first pricing parameter set.

[0053] S2, based on the weight set, perform a weighted operation on multiple first pricing parameters to obtain the tenant's second pricing parameters;

[0054] Specifically, a six-dimensional weight set for tenants is read from the caching system. This set includes cost weight, user weight, scenario weight, supply chain weight, profit weight, and time weight, all with a normalization constraint that the sum is 1. Each weight dimension is mapped one-to-one with its corresponding first pricing parameter to construct a matching matrix. The cost weight is directly assigned to the product value index, which comprehensively reflects the SKU's supply price, redemption efficiency, inventory decay, and expiration date, and is therefore related to the product's cost structure. The user weight is assigned to the user price sensitivity coefficient, which is derived through user behavior response modeling under price fluctuations, representing the user's sensitivity to price adjustments and measuring user acceptance and conversion potential. The scenario weight is assigned to the scenario conversion rate coefficient, which is determined by the target scenario (e.g., withdrawal, repayment, loan rejection page, etc.). The conversion rate over the historical period is compared to the platform average, reflecting the effectiveness of pricing incentives in the current usage scenario. Supply chain weights are assigned to the supply chain risk coefficient, which considers supplier price volatility and delivery fulfillment capabilities, reflecting the balance between cost fluctuations and supply assurance. Profit weights are allocated to the operational strategy weights, which are configured based on a combination of customer acquisition, user activation, and revenue generation strategies, and are related to profit-oriented behavior. Time weights are assigned to the holiday bonus coefficient, which is dynamically generated by combining cyclical forecasts of product sales with holiday characteristics to capture the price leverage effect brought about by peak demand. A weighted product model is constructed, multiplying each of the six parameters by its corresponding weight and summing them to form a one-dimensional fusion feature representation vector, serving as the tenant's second pricing parameter within the current pricing period.

[0055] S3, calculate the category benchmark price based on the product category label, and multiply the second pricing parameter by the category benchmark price to generate M first candidate prices;

[0056] It should be noted that the unique identifier of the target product is extracted from the main product table, and the category tag field bound to the target product is read to indicate the category to which the product belongs, such as rights, insurance, e-commerce, short drama, or notary. A routing table is constructed that maps category tags to mapping functions. This table predefines pricing factor calculation functions for five categories. For example, the `Φ_rights` function is used for rights-based products, calculating a reference price based on rights points, tenant point unit price, and discount rate. For insurance-based products, the `Φ_insure` function is called to generate a theoretical price for medical or accident insurance based on the insured amount, base rate, and user risk adjustment factor. For e-commerce products, the `Φ_mall` function combines supplier costs, target gross profit margin, and brand premium to form a standard selling price. For short dramas, the `Φ_drama` function calculates the cost of unlocking the entire series based on the number of episodes and the unlock price per episode. For notary-based products, the `Φ_notary` function is used to form a standardized fee range based on the notary amount, rate, and upper and lower thresholds. Once the category tag field of a product is identified, the corresponding pricing factor calculation function is called through the routing table. Then, the key parameter fields required for that category are extracted from the relevant data table to calculate the product's category benchmark price. The second pricing parameter is multiplied by the category benchmark price to form the original candidate price, representing the preferred theoretical price under the interplay of current tenant preferences, user scenarios, and product attributes. To cover different price-sensitive groups and conversion ranges, a preset price adjustment range set Δ={-0.2, -0.1, 0, 0.1,0.2} is introduced based on the benchmark price. M first candidate prices are generated through multiplication, forming a symmetrical price disturbance band with the pricing center as the base point.

[0057] S4. Verify each of the M first candidate prices, filter out the N second candidate prices that satisfy all constraints, calculate the comprehensive score of each second candidate price, and select the second candidate price with the highest comprehensive score as the target price and store it in the tenant pricing matrix.

[0058] Specifically, each candidate price is iterated sequentially, and each of the preset multi-dimensional constraints is verified, including cost lower bound constraints, profit target constraints, inventory capacity constraints, and market competitiveness constraints. For each candidate price, it is verified whether it is greater than or equal to the cost lower bound obtained by dividing the supplier's quote by the channel commission rate, to ensure that the tenant can still cover basic supply costs after paying channel costs. The product of the candidate price and its corresponding expected conversion rate and predicted exposure is calculated, and it is determined whether this value is greater than or equal to the profit target set by the tenant, to verify whether the price has sufficient profitability. It is assessed whether the expected sales volume corresponding to the candidate price under the current conversion rate and predicted exposure does not exceed the current inventory quantity; the inventory constraint is used to prevent overselling risk. The absolute value of the difference between the candidate price and the real-time competitor price is calculated and compared with the ratio of the competitor price. If the ratio is not greater than the preset competition deviation threshold, the price is considered to have a certain degree of market competitiveness. Candidate prices that satisfy all constraints are added to the second candidate price set. For each of the N candidate prices, a comprehensive score is calculated. This comprehensive score consists of three parts: expected revenue (measuring the commercial value of the pricing scheme), a predicted refund rate (one minus the corresponding refund rate, used to measure service stability after user acceptance of the price), and a user satisfaction index (constructed based on the deviation between the candidate price and the user's expected price; smaller deviations indicate higher satisfaction). These three scoring dimensions are multiplied by their respective weighting coefficients and summed to obtain the comprehensive score for each candidate price. All candidate prices are iterated through, and the one with the highest comprehensive score is selected as the final target price. This target price is then written into the tenant pricing matrix using a three-dimensional index composed of tenant identifier, product category identifier, and scenario identifier.

[0059] In one example, a weighted set and several primary pricing parameters are calculated based on each tenant's historical pricing records in the rights platform, including:

[0060] Obtain the historical pricing records for each tenant on the rights and interests platform. The historical pricing records include the face value of the product, the actual selling price, the supplier's cost, and the user's discount rate.

[0061] Calculate the ratio of supplier cost to actual selling price to obtain the cost percentage, and then calculate the difference between actual selling price and supplier cost and divide it by the actual selling price to obtain the profit percentage.

[0062] The cost weight is obtained by averaging the cost percentages and the profit weight is obtained by averaging the profit percentages. Default values ​​are set for supply chain weight, user weight, scenario weight, and time weight based on the tenant's business type.

[0063] Construct a weight set by considering cost weight, profit weight, supply chain weight, user weight, scenario weight, and time weight;

[0064] Calculate the user price sensitivity coefficient, product value index, scenario conversion rate coefficient, supply chain risk coefficient, operation strategy weight, and holiday bonus coefficient to generate multiple primary pricing parameters.

[0065] In this example, the historical pricing records for each tenant are retrieved from the platform's data warehouse. These records contain information such as the product's face value, actual selling price, supplier cost, and user discount rate for each item. The data spans the past 90 consecutive days to ensure the representativeness of the statistical sample. Each pricing record is structured and parsed, extracting the supplier cost and actual selling price fields. The cost percentage for each record is calculated by dividing the supplier cost by the actual selling price. The profit percentage is then calculated by dividing the difference between the actual selling price and the supplier cost by the actual selling price. The average of all cost percentages is used to calculate the tenant's cost weight. For example, if a tenant's average cost percentage over the 90-day pricing record is 0.4, then their cost weight is 0.4. Similarly, the average of all profit percentages is used to obtain the profit weight. If a tenant's average profit percentage over the statistical period is 0.35, then their profit weight is set to 0.35. The remaining four weights are assigned default values ​​based on the tenant's business focus. For example, if a tenant's primary business goal is user acquisition, its user weight is set to 0.25; if its advertising channels are concentrated in promotional scenarios, the scenario weight is set to 0.2; and if the suppliers it's affiliated with are long-tail suppliers with high volatility, the supply chain weight is set to 0.12. The time weight is set based on whether the tenant has holiday sales cycles, typically defaulting to a range of 0.08 to 0.1. These six weights are combined in a fixed order to form a unified weight set. Simultaneously, several primary pricing parameters are generated, including user price sensitivity coefficient, product value index, scenario conversion rate coefficient, supply chain risk coefficient, operational strategy weight, and holiday bonus coefficient. The user price sensitivity coefficient is calculated by analyzing the relationship between the price fluctuation and average of user orders over 180 days. The lower the price fluctuation, the more sensitive the user is to price. The commodity value index is a multi-factor index that reflects the quality of commodity assets by combining supplier quotations, redemption rate, inventory quantity, and validity period. If a commodity has a redemption rate of 0.85, an inventory of 1,000 units, and a validity period of 365 days, the commodity value index is higher than the threshold. The scenario conversion rate coefficient is calculated by the ratio between the current scenario payment conversion rate and the platform's average conversion rate. For example, if the cash withdrawal scenario conversion rate is 0. 4. A platform average of 0.25 results in a scenario conversion rate coefficient of 1.6, indicating excellent scenario conversion. The supply chain risk coefficient is assessed based on price volatility and on-time delivery rate; when price volatility is low and the delivery rate is high, the risk coefficient approaches 0. The operational strategy weights are assigned based on the tenant's configuration of new user acquisition, user activation, and revenue generation ratios. For example, a configuration of 0.3, 0.4, and 0.3 directly generates a three-way weighting. The holiday bonus coefficient is set based on whether the following week includes statutory holidays; for example, 0.5 for the Spring Festival, 0.2 for short holidays, and 0 for ordinary working days. These parameters together constitute multiple sets of first pricing parameters.

[0066] In one example, the user price sensitivity coefficient, product value index, scenario conversion rate coefficient, supply chain risk coefficient, operational strategy weight, and holiday bonus coefficient are calculated to generate multiple primary pricing parameters, including:

[0067] The price sensitivity coefficient of users is obtained by calculating the price variance and order frequency of users' historical orders from the user behavior record table. The commodity value index is calculated by extracting supplier quotations, write-off rates, inventory quantities and remaining validity periods from the equity SKU management table. The scenario conversion rate coefficient is calculated by statistically analyzing the historical conversion rates of cash withdrawal scenarios, repayment scenarios and loan rejection page scenarios from the order transaction table. The supply chain risk coefficient is calculated by extracting quotation volatility and supply stability from the supply chain management table. The operation strategy weight is read from the tenant configuration table to obtain the holiday bonus coefficient.

[0068] The pricing parameters include user price sensitivity coefficient, product value index, scenario conversion rate coefficient, supply chain risk coefficient, operational strategy weight, and holiday bonus coefficient.

[0069] In this example, all order data for the target user over the past 180 days is extracted from the user behavior log table. Within the time window, the prices of each order are aggregated to form a price series, and the mean and variance of this data set are calculated. The price mean reflects the user's central consumption level, while the price variance measures the degree of fluctuation in order prices. Combined with the number of orders *m* over the 180 days, the user's average order frequency, i.e., the frequency of orders placed per day, is calculated. The ratio between the price variance and the price mean is inversely transformed to obtain a user price sensitivity coefficient: the smaller the price fluctuation, the higher the sensitivity; the larger the fluctuation, the lower the sensitivity. For example, if the user's consumption price fluctuation is small, the sensitivity coefficient is close to 1, indicating that the user is highly sensitive to price changes; if the fluctuation is large, the sensitivity coefficient approaches 0.5, indicating that the user is more concerned about the product itself than its price. Extract core metrics for target SKUs from the equity SKU management table, including supplier quotes (e.g., 28 yuan), redemption rate in the past 30 days (e.g., 0.85), inventory quantity (e.g., 1000 units), and remaining shelf life (e.g., 365 days). Combine these four metrics to form a commodity value index. The inventory item decays by subtracting an exponential function from 1 to avoid excessive amplification of the overall value when there is a large inventory. For example, when the inventory is 1000 units, the item is about 0.63, while it approaches 1 when the inventory exceeds 5000 units. The remaining shelf life is annualized, which makes the value of goods that are about to expire significantly depreciate. The higher the index, the stronger the transaction value of the goods in the current period. The historical conversion performance of cash withdrawal, repayment, and loan rejection page scenarios is analyzed from the order transaction table. The number of payment orders and exposures for each scenario over the past 30 days are extracted, and their historical conversion rates are calculated. For example, the cash withdrawal scenario has a conversion rate of 0.4, while the platform average is 0.25. Therefore, the conversion rate coefficient for the cash withdrawal scenario is 1.6, reflecting that this entry point has a higher influence on user decisions than the platform average. The conversion rate coefficient is used to measure whether price adjustments have an amplifying or suppressing effect in this scenario. Simultaneously, the daily price sequence of SKUs over the past 14 days is read from the supply chain management table, and its volatility (e.g., standard deviation 0.1) is analyzed. This is combined with the percentage of on-time delivered orders from the supply fulfillment table (e.g., 0.9). Multiplying these two values ​​yields the supply chain risk coefficient. When price volatility is low and fulfillment rate is high, the risk coefficient is extremely low (e.g., 0.01). However, when prices fluctuate frequently or fulfillment is unstable, the risk coefficient rises rapidly, reminding the model to shift prices towards a more conservative risk direction. For the operational strategy weights, the platform-configured three-element strategy ratios are directly read from the tenant configuration table: new user acquisition weight, user activation weight, and revenue generation weight. For example, if configured as 0.3, 0.4, and 0.3, the weights are used to evaluate the impact of the tenant's profit-oriented goals on pricing.The system calls upon the holiday calendar to identify whether the pricing generation date is a regular day, a short holiday, or a major statutory holiday, and reads the corresponding holiday bonus coefficient. For example, a bonus of 0 for a regular day, 0.2 for a short holiday, and 0.5 for holidays like Spring Festival and National Day indicate a surge in demand during the holiday period, thus appropriately increasing the weight of the pricing strategy. These six parameters are then provided as the first pricing parameter group to the weighted calculation module.

[0070] In one example, a weighted operation is performed on multiple first pricing parameters based on a weight set to obtain the tenant's second pricing parameters, including:

[0071] Read the tenant's weight set, assign the cost weight in the weight set to the commodity value index among multiple first pricing parameters, assign the user weight to the user price sensitivity coefficient, assign the scenario weight to the scenario conversion rate coefficient, assign the supply chain weight to the supply chain risk coefficient, assign the profit weight to the operation strategy weight, and assign the time weight to the holiday bonus coefficient. Perform a weighted calculation to obtain the tenant's second pricing parameters.

[0072] In this example, the corresponding six-dimensional weight set is read from the tenant's cache space. This weight set consists of cost weight, user weight, scenario weight, supply chain weight, profit weight, and time weight, with a sum of 1. The weight values ​​are then assigned to the corresponding first pricing parameter dimension according to semantic relationships. The cost weight is directly mapped to the product value index, which is calculated from supplier quotes, write-off rates, inventory quantity, and expiration dates, among other cost and supply-demand related factors. This directly reflects the platform's perception of cost weights in the product dimension. The user weight is assigned to the user price sensitivity coefficient, a behavioral indicator determined by the user's historical order price fluctuations and activity level. This coefficient reflects the user's responsiveness to price adjustments. A higher user weight indicates that the tenant highly values ​​user acquisition or retention, leading to a higher proportion of such user sensitivity in the overall value. The scenario weight is mapped to the scenario conversion rate coefficient, calculated based on actual payment conversion data from target business entry points such as withdrawal, repayment, or loan rejection pages. This coefficient characterizes the contribution of different scenarios in driving conversion. If a tenant's business... Relying on specific scenario entry points amplifies the pricing effect of scenario weights. Supply chain weights are allocated to the supply chain risk coefficient, which measures the overall relationship between supplier price volatility and contract fulfillment stability. If a tenant has historically been frequently affected by unstable supply, this dimension's weight is increased to enhance protection against risky pricing. Profit weights are mapped to operational strategy weights, which exist in the first pricing parameter as a tripartite of customer acquisition, user activation, and revenue generation, representing the tenant's current profit-oriented strategy. A high profit weight indicates the tenant is currently in a period of strong profit targets and tends to push this parameter higher to support a higher price strategy. Time weights are allocated to holiday bonus coefficients, determined by the type of holiday (e.g., ordinary days, short holidays, or statutory holidays), responding to adjustments in demand cycles. A higher time weight indicates high sensitivity to seasonal fluctuations, resulting in a higher weighting for holiday bonus coefficients. The weighted calculation process multiplies each first pricing parameter by its corresponding weight, summing the six products to form a unified scalar value as the merged second pricing parameter.

[0073] In one example, a category benchmark price is calculated based on the product's category label, and a second pricing parameter is multiplied by the category benchmark price to generate M first candidate prices, including:

[0074] Query the category tags of the target product from the main product table, determine whether the product is a rights, insurance, mall, short drama or notary product based on the category tags, and construct a routing relationship table between the category tags and the corresponding mapping functions;

[0075] Select the mapping function corresponding to the category tag based on the routing table, and calculate the category benchmark price of the product based on the category pricing parameters corresponding to the mapping function;

[0076] The second pricing parameter is multiplied by the category benchmark price, and M first candidate prices are generated based on the category benchmark price by a preset adjustment range.

[0077] In this example, the product's unique identifier is used to retrieve its associated category label field from the main product table. The category label field, identified by `category_label`, indicates which core business category the product belongs to, such as "Rights," "Insurance," "Mall," "Short Drama," or "Notary." After identifying the category label, the binding relationship between the category and the mapping function is completed according to the category routing mapping table. The routing table maps each category label to its corresponding standardized benchmark pricing function. For example, rights products are bound to the `Φ_rights` function, whose core parameters are rights points, point unit price, and tenant discount rate; insurance products are mapped to the `Φ_insure` function, which calculates the theoretical price based on the insurance coverage, base rate, and user risk adjustment factor; mall products use the `Φ_mall` function, which models the product price using supplier costs, target gross profit margin, and brand premium factor; short drama products are bound to the `Φ_drama` function, whose calculation parameters are the total number of episodes and the unlock price per episode; and notary products correspond to the `Φ_notary` function, which generates a standard fee range based on the notarized amount and fixed rate combined with upper and lower limits. The system automatically selects the Φ_category function corresponding to the product based on the routing logic of the mapping table, and extracts the category pricing parameters that the function depends on from different business modules, such as reading costs from the supply chain system, gross profit margin or discount rate from tenant configuration, and episode count or notarized amount from the content management platform. The function is then executed to calculate the product's category benchmark price. The category benchmark price is multiplied by a second pricing parameter to form the tenant's original base price for the product, which serves as the central point of the candidate price range. A symmetrical price fluctuation range is set based on the category benchmark price, and a set of price adjustment ranges is defined, such as {-0.2, -0.1, 0, 0.1, 0.2}. Amplitude factors are applied to the category benchmark price to form upward or downward price adjustments, constructing M first candidate prices.

[0078] In one example, the mapping function corresponding to the category tag is selected based on the routing table, and the category base price of the product is calculated based on the category pricing parameters corresponding to the mapping function, including:

[0079] The mapping function corresponding to the category tag is retrieved from the routing table. For the rights category, it is mapped to the point multiplication function; for the insurance category, it is mapped to the coverage multiplication function; for the mall category, it is mapped to the cost division function; for the short drama category, it is mapped to the episode multiplication function; and for the notary category, it is mapped to the amount multiplication function.

[0080] The mapping function is called to substitute the corresponding category pricing parameters and perform the calculation to obtain the category benchmark price of the product.

[0081] In this example, a routing table is established to map category labels to mapping functions. This table maps the `category_label` field of each product to a predefined category calculation model function. For example, equity-type products are mapped to a points multiplication function, which takes the product's equity points, tenant point unit price, and tenant discount rate as input parameters. It then performs a combination of points × unit price × (1 - discount rate) to generate a reference price. Similarly, insurance-type products are mapped to a coverage multiplication function, which takes the insurance product's coverage amount, actuarial rate, and individual risk adjustment factor as core parameters. It then performs a multiplication of coverage amount × base rate × (1 + risk coefficient) to obtain the insurance theoretical value. Selling price; For mall-type products, the cost division function is used. The function structure is based on the supplier's quotation, the tenant's target gross profit margin, and the brand premium coefficient. By dividing the cost by (1 - gross profit margin) and then multiplying by the brand premium coefficient, a standard retail price model for traditional product SKUs is constructed. Short drama products are mapped to an episode multiplication function. The logic is to multiply the total number of episodes by the unlock price per episode. For example, 80 episodes multiplied by 0.12375 yuan / episode yields a total price of 9.9 yuan. Notary products are mapped to an amount multiplication function. The amount multiplication function structure multiplies the notary amount applied for by the user by a fixed rate and then determines whether the result falls within the allowed pricing range. The excess value is truncated according to the upper and lower limit rules to ensure pricing compliance. Based on the category tag field of the target product, if "insurance" is detected, the system routes to the Φ_insure function; if "drama" is detected, it routes to the Φ_drama function. The core calculation logic of the category mapping function is then invoked to extract necessary input parameters from the associated data tables. For example, benefit points are extracted from the main product table; point unit price and discount rate are obtained from the tenant configuration table; insurance coverage is extracted from the product field; premium rate and risk coefficient are jointly calculated from the actuarial table and user profile; mall costs are read from the supply chain system; brand premium is obtained from the brand level table; total number of episodes is retrieved from the content information table; single episode price is obtained from the pricing strategy table; and notarized amount is obtained from order application records. The extracted pricing parameters are automatically substituted into the matched mapping function for fixed-value calculations, resulting in a standardized, structured product benchmark price that conforms to category differences. For example, when an insurance product is identified with a coverage of 6 million yuan, a premium rate of 0.0005, and a user risk coefficient of 0.1, the benchmark price is directly calculated as 3300 yuan through the mapping function.

[0082] In one example, the second pricing parameter is multiplied by the category benchmark price, and M first candidate prices are generated based on the category benchmark price by a preset adjustment range, including:

[0083] The base price is obtained by multiplying the second pricing parameter by the category benchmark price;

[0084] The difference between the scene conversion rate coefficient and the platform average conversion rate is calculated, divided by the platform average conversion rate, multiplied by the preset scene sensitivity parameter, and then one is added to obtain the scene adjustment coefficient.

[0085] Multiply the base value by the scenario adjustment coefficient to obtain the scenario-adjusted price;

[0086] Extract the price change rate and sales change rate of tenants under the product from the order history database. Divide the sales change rate by the price change rate and take the average to obtain the price elasticity coefficient. Multiply the scenario-adjusted price by the correction factor corresponding to the price elasticity coefficient to obtain the elasticity correction price.

[0087] Based on the elastic adjustment price, M first candidate prices are calculated according to the preset adjustment range.

[0088] In this example, the second pricing parameter, which integrates multiple factors, is multiplied by the category benchmark price to obtain the base pricing value. This base pricing value serves as the initial benchmark for the theoretical price, reflecting the combined output of tenant pricing preferences and product category value. A dynamic scenario adjustment mechanism is introduced. The conversion performance of the current business scenario (e.g., cash withdrawal) is extracted from the scenario conversion rate coefficient, and the average conversion rate is calculated from all scenarios across the platform. For example, if the current scenario conversion rate is 0.4 and the platform average is 0.25, the difference is divided by the average to obtain the relative increase ratio. This ratio is then multiplied by a preset scenario sensitivity parameter, for example, set to 0.3, to control the adjustment magnitude. One is added to this result to obtain the scenario adjustment coefficient. For example, when the difference is 0.15, the adjustment coefficient is 1.18, indicating that the original price needs to be increased in this scenario. Multiplying the adjustment coefficient by the base pricing value yields the scenario-adjusted price, reflecting the system's dynamic adaptation strategy to conversion efficiency under the current business entry point. A price elasticity adjustment mechanism is introduced. It retrieves the tenant's actual transaction data for the product over the past 30 days from the order history database, extracting the price and sales volume changes for each two-day period. The sales volume change rate is divided by the corresponding price change rate for all intervals to calculate a series of elasticity ratios. The average of these elasticity ratios yields the price elasticity coefficient. For example, if a 10% price increase leads to a 15% decrease in sales, the average elasticity coefficient is -1.5. A correction factor is then matched based on the tenant's pricing strategy (e.g., 0.2 for aggressive strategies and 0.5 for conservative strategies). The adjusted price is multiplied by the product of the correction factor and the elasticity coefficient to obtain the adjusted price. Using this adjusted price as the center, a price perturbation generation module is invoked. A preset set of adjustment ranges (e.g., -20%, -10%, 0%, +10%, +20%) is applied to this price, creating upper and lower fluctuation bands. M first-choice prices are then calculated.

[0089] In one example, M first-choice prices are validated, N second-choice prices that satisfy all constraints are selected, a comprehensive score is calculated for each second-choice price, and the second-choice price with the highest comprehensive score is selected as the target price and stored in the tenant pricing matrix, including:

[0090] For each of the M first candidate prices, verify whether they are greater than or equal to the lower bound of the cost difference between the supplier's price and the channel commission ratio, whether the product of the candidate price and the expected conversion rate and the predicted exposure is greater than or equal to the tenant's profit target, whether the expected sales volume is less than or equal to the inventory quantity, and whether the absolute value of the difference between the candidate price and the competitor's price divided by the competitor's price is less than or equal to the preset deviation threshold. Select N second candidate prices that simultaneously satisfy all constraints.

[0091] For N second candidate prices, calculate the expected revenue, predicted refund rate, and user satisfaction. Then, perform a weighted summation of the expected revenue, the predicted refund rate, and user satisfaction to obtain the comprehensive score for each second candidate price.

[0092] Iterate through N second-candidate prices and select the one with the highest comprehensive score as the target price. Store the target price in the tenant pricing matrix using the tenant identifier, product category identifier, and scenario identifier as index keys.

[0093] In this example, each candidate price is subjected to four constraints according to a preset logic. The first constraint is the cost lower bound. The supplier's quote for the current product is retrieved from the supply chain system, for example, 28 yuan. Combined with the channel commission rate of 0.05 read from the tenant configuration table, the lower limit of the price is calculated as 28 divided by (1 minus 0.05), resulting in 29.47 yuan. This checks whether the candidate price is not lower than the lower limit to ensure that the tenant still has a basic profit margin after covering costs and channel fees. The second constraint is the profit target constraint. The candidate price is multiplied by the expected conversion rate (e.g., 0.28) and the predicted number of users exposed (e.g., 50,000), and then multiplied by the pricing period (e.g., 7 days). If the result is not less than the tenant's profit target (e.g., 300,000 yuan), the profitability requirement is considered met. An inventory constraint judgment is then performed: the expected sales volume calculated from the candidate price, conversion rate, and exposure is compared with the inventory quantity (e.g., 8,000 units). If the sales volume is less than or equal to the inventory, the inventory risk is controllable. A competitiveness constraint is then applied: the current selling price of competing products (e.g., 75 yuan) is captured, and the absolute value of the difference between the candidate price and the competitor's price is compared. This difference is then divided by the competitor's price to determine if the deviation ratio is less than a preset threshold (e.g., 15%), ensuring the price remains competitive in the market. After these four verifications, candidate prices that meet all conditions are retained to form a second candidate price set. For each price in the second candidate price set, its expected revenue, predicted refund rate, and user satisfaction are calculated sequentially as scoring criteria. Expected revenue is calculated directly based on candidate price, conversion rate, and exposure. For example, with a price of 79.4 yuan, a conversion rate of 0.28, and 50,000 exposures, the expected revenue reaches 1,111,600 yuan. The predicted refund rate is derived by the system using a logistic regression model fitted to historical refund data near the given price. Assuming a rate of 0.08, its complement value is 0.92, indicating higher service quality. User satisfaction is assessed based on the deviation between the current candidate price and the user's psychological price expectations from survey data. For example, with an expected mean of 72 yuan and a standard deviation of 8 yuan, a candidate price of 79.4 yuan results in an estimated satisfaction score of 81.5, indicating the price is within an acceptable range. The three indicators are weighted and summed according to their respective weights. For example, revenue has a weight of 0.5, the refund rate complement value has a weight of 0.3, and satisfaction has a weight of 0.2, resulting in a comprehensive score. Iterate through all N second candidate prices, select the one with the highest comprehensive score as the final target price, and write the target price into the tenant pricing matrix using the three-element combination of tenant identifier, product category identifier and business scenario identifier as the index key.

[0094] In one example, for M first candidate prices, the following checks are performed sequentially: whether they are greater than or equal to the lower bound of the cost difference between the supplier's quoted price and the channel commission rate; whether the product of the candidate price and the expected conversion rate and predicted exposure is greater than or equal to the tenant's profit target; whether the expected sales volume is less than or equal to the inventory quantity; and whether the absolute value of the difference between the candidate price and the competitor's price divided by the competitor's price is less than or equal to a preset deviation threshold. N second candidate prices that simultaneously satisfy all constraints are then selected, including:

[0095] Read supplier quotes from the equity SKU management table and channel commission rates from the tenant configuration table. Calculate the difference between supplier quotes and channel commission rates to obtain the lower limit of costs. Read tenant profit targets from the tenant configuration table and inventory quantities from the inventory management table. Obtain competitor prices and preset deviation thresholds from the competitor monitoring system.

[0096] For each of the M first candidate prices, verify whether it is greater than or equal to the lower bound of the cost difference between the supplier's price and the channel commission rate, whether the product of the candidate price and the expected conversion rate and the predicted exposure is greater than or equal to the tenant's profit target, whether the expected sales volume is less than or equal to the inventory quantity, and whether the absolute value of the difference between the candidate price and the competitor's price divided by the competitor's price is less than or equal to the preset deviation threshold. Select N second candidate prices that simultaneously satisfy all constraints.

[0097] In this example, the supplier quote for the target product, for example, 28 yuan, is extracted from the equity SKU management table. Simultaneously, the channel commission ratio set by the corresponding tenant, for example, 0.05, is extracted from the tenant configuration table. The lowest acceptable price of the product, i.e., the lower bound of cost, is calculated by dividing the supplier quote by (1 minus the channel commission ratio), resulting in 29.47 yuan, which serves as the first constraint boundary for price reasonableness verification. The profit target value set by the tenant, assumed to be 300,000 yuan, is read from the tenant configuration table and recorded as a core reference indicator for profitability determination. The current inventory quantity, for example, 8,000 units, is read from the inventory management table as the basis for judging sales capacity constraints. The current market sales price of the competitor corresponding to the target product, for example, 75 yuan, is retrieved from the competitor monitoring system, and the set price deviation tolerance threshold, for example, θ of 15%, is simultaneously obtained to assess whether the price has sufficient market competitiveness. For each of the M first-choice prices, constraints are sequentially verified. For each price, it is determined whether it is greater than or equal to the lower cost limit of 29.47 yuan; if it is, it is directly eliminated. Each price is then multiplied by its corresponding expected conversion rate (e.g., 0.28), predicted exposure (e.g., 50,000), and pricing period (e.g., 7 days). If the result is greater than or equal to the profit target of 300,000 yuan, the price is considered profitable. Next, it is determined whether the expected sales volume corresponding to the price under the given exposure and conversion rate is less than or equal to the inventory quantity of 8,000 units to avoid triggering overselling risk. The absolute value of the difference between the price and the 75 yuan competitor price is compared, and the difference is divided by 75 yuan. If the resulting percentage is less than or equal to 15%, the price is still within a reasonable market competition range; otherwise, it is considered a price that deviates too much. All candidate prices that meet the above four constraints are uniformly included in the compliant price set, forming N second-choice prices.

[0098] Reference Figure 2 This embodiment provides an intelligent product pricing strategy generation system that integrates multi-factor analysis, including:

[0099] Pricing module 1 is used to calculate a weight set and multiple first pricing parameters based on the historical pricing records of each tenant in the rights platform;

[0100] Calculation module 2 is used to perform a weighted calculation on multiple first pricing parameters according to the weight set to obtain the tenant's second pricing parameters;

[0101] The benchmark price calculation module 3 is used to calculate the category benchmark price based on the product category label, and multiply the second pricing parameter by the category benchmark price to generate M first candidate prices;

[0102] The filtering module 4 is used to verify the M first candidate prices, filter the N second candidate prices that meet all constraints, calculate the comprehensive score of each second candidate price, and select the second candidate price with the highest comprehensive score as the target price and store it in the tenant pricing matrix.

[0103] In this embodiment, the specific implementation of each unit in the above system embodiment is described in the above method embodiment, and will not be repeated here.

[0104] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, system, article, or method that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, system, article, or method. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, system, article, or method that includes that element.

[0105] The above description is only a preferred embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structural or procedural transformations made based on the content of the present invention specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. An intelligent commodity pricing strategy generation method based on fusion multi-factor analysis, characterized in that, The method comprises the following steps: calculating a weight set and a plurality of first pricing parameters according to historical pricing records of each tenant in the benefit platform; Specifically, the method comprises the following steps: obtaining historical pricing records of each tenant in the benefit platform, wherein the historical pricing records include a commodity face value, an actual selling price, a supplier cost, and a user discount rate; calculating a ratio of the supplier cost to the actual selling price to obtain a cost proportion, and calculating a difference between the actual selling price and the supplier cost divided by the actual selling price to obtain a profit proportion; calculating an average value of the cost proportion to obtain a cost weight, and calculating an average value of the profit proportion to obtain a profit weight; setting default values for a supply chain weight, a user weight, a scenario weight, and a time weight according to a business type of the tenant; constructing a weight set from the cost weight, the profit weight, the supply chain weight, the user weight, the scenario weight, and the time weight; calculating a user price sensitivity coefficient, a commodity value index, a scenario conversion rate coefficient, a supply chain risk coefficient, an operation strategy weight, and a holiday bonus coefficient to generate a plurality of first pricing parameters; performing weighted operation on the plurality of first pricing parameters according to the weight set to obtain a second pricing parameter of the tenant; calculating a category benchmark price according to a category label of a commodity, and multiplying the second pricing parameter by the category benchmark price to generate M first candidate prices; verifying the M first candidate prices respectively, screening N second candidate prices that meet all constraints, calculating a comprehensive score of each second candidate price, and selecting a second candidate price with the highest comprehensive score as a target price and storing the target price in a tenant pricing matrix. 2.The intelligent commodity pricing strategy generation method of fusion multi-factor analysis according to claim 1, characterized in that, The method of calculating a user price sensitivity coefficient, a commodity value index, a scenario conversion rate coefficient, a supply chain risk coefficient, an operation strategy weight, and a holiday bonus coefficient to generate a plurality of first pricing parameters comprises the following steps: calculating a user price sensitivity coefficient from a price variance and order frequency of a user historical order in a user behavior record table, calculating a commodity value index from a supplier quote, a cancellation rate, a stock quantity, and a remaining valid period in a benefit SKU management table, calculating a scenario conversion rate coefficient from historical conversion rates of a scenario, a repayment scenario, and a loan rejection page scenario in an order transaction table, calculating a supply chain risk coefficient from a quote fluctuation rate and a supply stability in a supply chain management table, reading an operation strategy weight from a tenant configuration table, and obtaining a holiday bonus coefficient; The user price sensitivity coefficient, the commodity value index, the scenario conversion rate coefficient, the supply chain risk coefficient, the operation strategy weight, and the holiday bonus coefficient are used as a plurality of first pricing parameters. 3.The intelligent commodity pricing strategy generation method of fusion multi-factor analysis according to claim 2, characterized in that, The method of performing weighted operation on the plurality of first pricing parameters according to the weight set to obtain a second pricing parameter of the tenant comprises the following steps: reading a weight set of the tenant, assigning a cost weight in the weight set to a commodity value index in the plurality of first pricing parameters, assigning a user weight to the user price sensitivity coefficient, assigning a scene weight to the scene conversion rate coefficient, assigning a supply chain weight to the supply chain risk coefficient, assigning a profit weight to the operation strategy weight, and assigning a time weight to the holiday addition coefficient, performing a weighted operation to obtain second pricing parameters of the tenant. 4.The intelligent commodity pricing strategy generation method of fusion multi-factor analysis according to claim 1, characterized in that, The second pricing parameters are multiplied by a category benchmark price of the commodity to generate M first candidate prices, including: querying a category label of the target commodity from a commodity master table, determining that the commodity is a benefit type, an insurance type, a mall type, a short drama type, or a notarization type according to the category label, and constructing a routing relationship table of the category label and a corresponding mapping function; selecting a mapping function corresponding to the category label according to the routing relationship table, and calculating a category benchmark price of the commodity based on a category pricing parameter corresponding to the mapping function; multiplying the second pricing parameters by the category benchmark price and generating M first candidate prices at a preset adjustment amplitude based on the category benchmark price. 5.The intelligent commodity pricing strategy generation method of fusion multi-factor analysis according to claim 4, characterized in that, The second pricing parameters are multiplied by a category benchmark price of the commodity to generate M first candidate prices, including: querying a mapping function corresponding to the category label from the routing relationship table, mapping a point multiplication function for the benefit type, mapping a premium multiplication function for the insurance type, mapping a cost division function for the mall type, mapping a number of episodes multiplication function for the short drama type, and mapping an amount multiplication function for the notarization type; calling the mapping function to substitute the corresponding category pricing parameter into the mapping function to obtain the category benchmark price of the commodity. 6.The intelligent commodity pricing strategy generation method of fusion multi-factor analysis according to claim 4, characterized in that, The second pricing parameters are multiplied by a category benchmark price of the commodity to generate M first candidate prices at a preset adjustment amplitude based on the category benchmark price, including: multiplying the second pricing parameters by the category benchmark price to obtain a basic fixed value; calculating a difference between a scene conversion rate coefficient and a platform average conversion rate, dividing the difference by the platform average conversion rate, multiplying a preset scene sensitivity parameter, and adding one to obtain a scene adjustment coefficient; multiplying the basic fixed value by the scene adjustment coefficient to obtain a scene adjustment price; extracting a price change rate and a sales change rate of the tenant under the commodity from an order history database, calculating a price elasticity coefficient by dividing the sales change rate by the price change rate and taking an average value, multiplying the scene adjustment price by a correction factor corresponding to the price elasticity coefficient to obtain an elasticity correction price; based on the elasticity correction price, M first candidate prices are calculated at a preset adjustment amplitude. 7.The intelligent commodity pricing strategy generation method of fusion multi-factor analysis according to claim 1, wherein, The M first candidate prices are verified respectively, N second candidate prices satisfying all constraints are screened, a comprehensive score of each second candidate price is calculated, the second candidate price with the highest comprehensive score is selected as a target price, and the target price is stored in a tenant pricing matrix, including: verify whether the M first candidate prices are greater than or equal to a cost lower bound of a supplier price divided by a channel commission rate difference value, whether a product of a candidate price and an expected conversion rate and a predicted exposure amount is greater than or equal to a tenant profit target, whether an expected sales volume is less than or equal to a stock quantity, and whether an absolute value of a difference between the candidate price and a competitor price divided by the competitor price is less than or equal to a preset deviation threshold value, and filter N second candidate prices that simultaneously satisfy all the constraints; calculate an expected revenue, a predicted refund rate, and a user satisfaction degree for the N second candidate prices, perform a weighted summation on the expected revenue, a complement of the predicted refund rate, and the user satisfaction degree, and obtain a comprehensive score of each second candidate price; select a second candidate price with the highest comprehensive score as a target price by traversing the N second candidate prices, and store the target price in a tenant pricing matrix with a tenant identifier, a commodity category identifier, and a scenario identifier as index keys. 8.The intelligent commodity pricing strategy generation method of fusion multi-factor analysis according to claim 7, characterized in that, The verifying whether the M first candidate prices are greater than or equal to the cost lower bound of the supplier price divided by the channel commission rate difference value, whether the product of the candidate price and the expected conversion rate and the predicted exposure amount is greater than or equal to the tenant profit target, whether the expected sales volume is less than or equal to the stock quantity, and whether the absolute value of the difference between the candidate price and the competitor price divided by the competitor price is less than or equal to the preset deviation threshold value, and filtering the N second candidate prices that simultaneously satisfy all the constraints, comprise: reading the supplier price from an equity SKU management table and reading the channel commission rate from a tenant configuration table, calculating the cost lower bound of the supplier price divided by the channel commission rate difference value, reading the tenant profit target from the tenant configuration table, reading the stock quantity from a stock management table, and obtaining the competitor price and the preset deviation threshold value from a competitor monitoring system; The verifying whether the M first candidate prices are greater than or equal to the cost lower bound of the supplier price divided by the channel commission rate difference value, whether the product of the candidate price and the expected conversion rate and the predicted exposure amount is greater than or equal to the tenant profit target, whether the expected sales volume is less than or equal to the stock quantity, and whether the absolute value of the difference between the candidate price and the competitor price divided by the competitor price is less than or equal to the preset deviation threshold value, and filtering the N second candidate prices that simultaneously satisfy all the constraints.

9. An intelligent commodity pricing strategy generation system based on fusion multi-factor analysis, characterized in that, The steps for implementing the intelligent commodity pricing strategy generation method of the fusion multi-factor analysis in any one of claims 1 to 8, comprising: a pricing module configured to calculate a weight set and a plurality of first pricing parameters according to historical pricing records of each tenant in an equity platform; an operation module configured to perform a weighted operation on the plurality of first pricing parameters according to the weight set to obtain a second pricing parameter of the tenant; a benchmark price calculation module configured to calculate a category benchmark price according to a category label of a commodity, and multiply the second pricing parameter by the category benchmark price to generate M first candidate prices; a filtering module configured to verify the M first candidate prices respectively, filter N second candidate prices that satisfy all the constraints, calculate a comprehensive score of each second candidate price, and select a second candidate price with the highest comprehensive score as a target price and store it in a tenant pricing matrix.

Images