Brand marketing system and method based on block chain

Abstract

The invention relates to the technical field of brand marketing, in particular to a brand marketing system and method based on a block chain. Comprising a market monitoring module, a dynamic pricing module, a consumption prediction module and an evaluation optimization module. The market data is monitored in real time through the market monitoring module, a multivariate linear regression model between market demands and prices is established through the dynamic pricing module, the price of the brand is dynamically adjusted by taking profit maximization as a target, the consumption prediction module is utilized to construct a random forest model, and the price of the brand is dynamically adjusted. Each consumer is classified into a corresponding level according to a comprehensive score, personalized marketing suggestions are generated, the price of a brand is dynamically adjusted according to the change of market demands, the price of the brand can be ensured to conform to the change of the market in real time, the purpose of brand promotion is achieved, targeted marketing of consumer groups is achieved, and the market promotion efficiency is improved. And the marketing quality and efficiency of the brand are improved.

Description

Technical Field

[0001] The present invention relates to the field of brand marketing technology, and in particular to a blockchain-based brand marketing system and method thereof. Background Art

[0002] The blockchain-based brand marketing system brings new possibilities for the interaction between brands and consumers by improving transparency, enhancing data security, and achieving precision marketing and personalized services. With the continuous advancement of technology and the continuous expansion of application scenarios, blockchain technology is expected to become an important driving force in the field of brand marketing in the future, creating greater value for brands and consumers. Blockchain technology, as an important breakthrough in the field of science and technology in recent years, is gradually changing the operating mode of various industries. In the field of brand marketing, the application of blockchain technology not only improves the transparency and security of marketing activities, but also brings new possibilities for the interaction between brands and consumers.

[0003] At present, smart contracts on blockchain platforms can monitor market data in real time and record market data using distributed ledgers, but lack comprehensive analysis of key indicators of market supply and demand, competitive product prices, and consumer behavior. In order to dynamically adjust the price of the brand based on the market data recorded in real time by the blockchain platform, and use the brand's price adjustment records and consumer behavior data to stratify consumer groups and generate corresponding personalized marketing suggestions for the corresponding groups, we propose a blockchain-based brand marketing system and method. Summary of the invention

[0004] The purpose of the present invention is to solve the problem that the blockchain platform lacks comprehensive analysis of key indicators of market supply and demand, competitor prices and consumer behavior, in order to dynamically adjust the price of the brand according to the market data recorded in real time by the blockchain platform, and generate corresponding personalized marketing strategies for consumer groups based on the brand's price adjustment records and consumer behavior data.

[0005] To achieve the above-mentioned purpose, the present invention provides a brand marketing system based on blockchain, including a market monitoring module, a dynamic pricing module, a consumption prediction module and an evaluation optimization module;

[0006] The market monitoring module monitors market data in real time through smart contracts on the blockchain platform, and uses distributed ledgers to record key indicators of market supply and demand, competitive product prices, and consumer behavior, and cleans, converts, and integrates the collected data to generate standardized data sets, which are then passed to the dynamic pricing module and the consumption forecasting module;

[0007] The dynamic pricing module establishes a multiple linear regression model between market demand and price based on the market demand data and price data recorded by the market monitoring module, uses the multiple linear regression model to predict market demand at different prices, and dynamically adjusts the price of the brand with the goal of maximizing profits, and transmits the price adjustment record of the brand to the consumption prediction module;

[0008] The consumption prediction module constructs a random forest model based on the price adjustment data transmitted by the dynamic pricing module and the consumer behavior data transmitted by the market monitoring module to predict the consumer's future purchasing behavior and preferences, uses the hierarchical analysis method to determine the indicator weights, comprehensively scores the consumers, and sets the stratification threshold, classifies each consumer into the corresponding level according to the comprehensive score, and generates personalized marketing suggestions;

[0009] The evaluation and optimization module utilizes the market monitoring module to monitor the market data in real time, utilizes the stepwise regression method to optimize and update the variables of the multivariate linear regression model, and optimizes and adjusts the parameters of the random forest model by adjusting the sample sampling ratio and feature sampling ratio of the model.

[0010] Preferably, the dynamic pricing module includes a model building unit and a price adjustment unit;

[0011] The model building unit establishes a multivariate linear regression model based on the recorded market demand data and price data, calculates the mean square error between the actual observed value and the model predicted value, and uses the minimum mean square error value as the evaluation index of the model to predict the market demand under different prices;

[0012] The price adjustment unit calculates the profits under different price and demand combinations based on the predicted market demand and cost data, draws the profit-price curve and the profit-demand curve, observes the changing trend of profit with price and demand, and dynamically adjusts the price of the brand with the goal of maximizing profit.

[0013] Preferably, the formula for establishing the multivariate linear regression model by the model building unit is:

[0014] ;

[0015] in, To meet market demand, is the intercept, For price, For seasonal factors, For price The corresponding regression coefficient is Seasonal factors The corresponding regression coefficient is is the error term.

[0016] Preferably, the model building unit estimates the parameters of the multivariate linear regression model by the least squares method to minimize the sum of squared errors between actual observed values ​​and model predicted values ​​to determine the value of the regression coefficient.

[0017] Preferably, the price adjustment unit adopts cost-plus pricing method to determine the base price, and determines the triggering condition of price adjustment and the price adjustment range according to the ratio between the predicted value and the actual value of the multivariate linear regression model.

[0018] Preferably, the consumption prediction module includes a prediction model unit and a personalized marketing unit;

[0019] The prediction model unit divides the price adjustment data and the consumer behavior data into 70% training set and 30% test set, trains the random forest model using the training set data, adjusts the model parameters, and uses the test set to evaluate the accuracy of the model prediction;

[0020] The personalized marketing unit uses the hierarchical analysis method to determine the indicator weights, comprehensively score consumers, and set stratification thresholds, classifying each consumer into a corresponding level according to the comprehensive score to generate personalized marketing suggestions.

[0021] Preferably, the prediction model unit measures the degree of reduction of information uncertainty before and after splitting the node by information gain, and uses the minimum information gain threshold as the cutoff condition for splitting the random forest model.

[0022] Preferably, the personalized marketing unit provides personalized marketing suggestions to consumer groups at different levels, tests different marketing strategies using A / B testing, and determines the most effective marketing strategy by observing consumers' click-through rates and conversion rates.

[0023] Preferably, when the evaluation and optimization module optimizes and updates the variables of the multivariate linear regression model using the stepwise regression method, the steps of forward selection and backward selection are performed alternately. Forward selection is performed first, and an independent variable that improves the model the most is selected from the remaining independent variables that have not been added, and then backward selection is performed to check the independent variables that have been added to the model and eliminate the independent variable that contributes the least to the model.

[0024] The second object of the present invention is to provide a blockchain-based brand marketing method, including any one of the blockchain-based brand marketing systems described above, comprising the following steps:

[0025] S1. Use the market monitoring module to monitor market data in real time, and use the distributed ledger to record key indicators of market supply and demand, competitive product prices, and consumer behavior, and convert them into standardized data sets;

[0026] S2. Establish a multivariate linear regression model between market demand and price through the dynamic pricing module, predict the market demand under different prices, and dynamically adjust the price of the brand with the goal of maximizing profits;

[0027] S3. Use the consumption prediction module to build a random forest model to predict consumers' future purchasing behavior and preferences. Use the hierarchical analysis method to determine the indicator weights, conduct a comprehensive score on consumers, set the stratification threshold, classify each consumer into the corresponding level according to the comprehensive score, and generate personalized marketing suggestions.

[0028] S4. The variables of the multivariate linear regression model are optimized and updated by using the stepwise regression method through the evaluation and optimization module, and the parameters of the random forest model are optimized by adjusting the sample sampling ratio and feature sampling ratio of the random forest model.

[0029] Compared with the prior art, the present invention has the following beneficial effects:

[0030] 1. The blockchain-based brand marketing system and method utilizes a market monitoring module to monitor market data in real time, and utilizes a distributed ledger to record key indicators of market supply and demand, competitive product prices, and consumer behavior, and converts them into standardized data sets. A multivariate linear regression model between market demand and price is established through a dynamic pricing module to predict market demand at different prices, and the brand price is dynamically adjusted with the goal of maximizing profits, so as to dynamically adjust the brand price according to changes in market demand, ensure that the brand price can keep up with market changes in real time, and achieve the purpose of brand promotion.

[0031] 2. The consumption prediction module constructs a random forest model based on the price adjustment data transmitted by the dynamic pricing module and the consumer behavior data transmitted by the market monitoring module, stratifies the consumer groups, and classifies each consumer into the corresponding level according to the comprehensive score. It generates personalized marketing suggestions, associates the price adjustment of the brand with the consumer's consumption behavior, realizes targeted marketing to the consumer group, and improves the marketing quality and efficiency of the brand. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 It is the overall flow chart of the present invention;

[0033] Figure 2 It is the overall detailed flow chart of the present invention;

[0034] Figure 3 The present invention is a flow chart of the method.

[0035] The meaning of each number in the figure is:

[0036] 100. Market monitoring module; 200. Dynamic pricing module; 210. Model building unit; 220. Price adjustment unit; 300. Consumption forecasting module; 310. Forecasting model unit; 320. Personalized marketing unit; 400. Evaluation and optimization module. DETAILED DESCRIPTION

[0037] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0038] At present, the blockchain platform lacks comprehensive analysis of key indicators of market supply and demand, competitor prices and consumer behavior. In order to be able to dynamically adjust the brand's price based on the market data recorded in real time by the blockchain platform, and generate corresponding personalized marketing strategies for consumer groups based on the brand's price adjustment records and consumer behavior data.

[0039] Therefore, the present invention proposes to monitor market data in real time through a market monitoring module, establish a multivariate linear regression model between market demand and price through a dynamic pricing module, predict market demand at different prices, and dynamically adjust the price of the brand with the goal of maximizing profits, construct a random forest model using a consumption prediction module to predict consumers' future purchasing behavior and preferences, determine indicator weights using a hierarchical analysis method, comprehensively score consumers, set stratification thresholds, classify each consumer into a corresponding level according to the comprehensive score, generate personalized marketing recommendations, optimize and update the variables of the multivariate linear regression model using a stepwise regression method through an evaluation and optimization module, and optimize and adjust the parameters of the random forest model by adjusting the sample sampling ratio and feature sampling ratio of the random forest model.

[0040] The details are as follows:

[0041] like Figure 1 As shown, one of the purposes of the present invention is to provide a brand marketing system based on blockchain, including a market monitoring module 100, a dynamic pricing module 200, a consumption prediction module 300 and an evaluation optimization module 400;

[0042] The market monitoring module 100 monitors the market data in real time through the smart contracts on the blockchain platform, and uses the distributed ledger to record the key indicators of market supply and demand, competitive product prices and consumer behavior, and cleans, converts and integrates the collected data to generate a standardized data set, and transmits the data set to the dynamic pricing module 200 and the consumption prediction module 300;

[0043] The dynamic pricing module 200 establishes a multiple linear regression model between market demand and price based on the market demand data and price data recorded by the market monitoring module 100, uses the multiple linear regression model to predict market demand at different prices, and dynamically adjusts the price of the brand with the goal of maximizing profits, and transmits the price adjustment record of the brand to the consumption prediction module 300;

[0044] The consumption prediction module 300 constructs a random forest model based on the price adjustment data transmitted by the dynamic pricing module 200 and the consumer behavior data transmitted by the market monitoring module 100 to predict the consumer's future purchasing behavior and preferences, uses the hierarchical analysis method to determine the indicator weights, comprehensively scores the consumers, and sets the stratification threshold to classify each consumer into the corresponding level according to the comprehensive score, and generate personalized marketing suggestions;

[0045] The evaluation and optimization module 400 uses the market monitoring module 100 to monitor the market data in real time, uses the stepwise regression method to optimize and update the variables of the multivariate linear regression model, and optimizes and adjusts the parameters of the random forest model by adjusting the sample sampling ratio and feature sampling ratio;

[0046] The market monitoring module 100 collects market data from multiple channels such as e-commerce platforms, social media, search engines, etc. in real time, including consumer behavior data (such as purchase records, search records, browsing records, interaction records, etc.), competitor price data and market supply and demand data, etc., and cleans, converts and integrates the collected data to generate a standardized data set.

[0047] like Figure 2 As shown, the dynamic pricing module 200 includes a model building unit 210 and a price adjustment unit 220;

[0048] The model building unit 210 builds a multivariate linear regression model according to the recorded market demand data and price data, calculates the mean square error between the actual observed value and the model predicted value, and uses the minimum mean square error value as the evaluation index of the model to predict the market demand under different prices;

[0049] The price adjustment unit 220 calculates the profits under different price and demand combinations based on the predicted market demand and the cost data, and dynamically adjusts the price of the brand with the goal of maximizing profit by drawing a profit-price curve and a profit-demand curve to observe the changing trend of profit with price and demand;

[0050] Collect actual sales volume and sales amount of products over a period of time, and other indicators that can reflect market demand. Data sources can be internal sales records of the company, data released by market research institutions, etc., and record the sales price of products in the corresponding time period, including the actual transaction price under different promotional activities;

[0051] Check whether there are missing values, outliers, etc. in the data. For missing values, you can use methods such as deletion and interpolation (such as mean interpolation and median interpolation). For outliers, you need to analyze the reasons for their occurrence. If it is an input error, correct it. If it is indeed an outlier caused by special circumstances, you can consider retaining it or making appropriate transformations.

[0052] In order to better construct the multiple linear regression model, the formula for establishing the multiple linear regression model by the model building unit 210 is:

[0053] ;

[0054] in, To meet market demand, is the intercept, For price, For seasonal factors, For price The corresponding regression coefficient is Seasonal factors The corresponding regression coefficient is is the error term.

[0055] Suppose we are interested in the automobile market in a certain region and have collected sales data for a certain type of car in the region over the past six months, including monthly sales prices (unit: 10,000 yuan) and sales volumes (unit: vehicles). At the same time, considering that automobile demand may be affected by the economic environment, we also collected monthly average disposable income of residents (unit: yuan) as an additional independent variable. The specific data is as follows:

[0056] month Sales price (ten thousand yuan) Sales volume (units) 1 15 100 2 14.5 120 3 15.5 90 4 14 130 5 16 80 6 13.5 140

[0057] Use Python's sciki-learn library to train the model and estimate the regression coefficient. After running the code, the regression coefficient and intercept are as follows: The sales price coefficient is 560.34. The average disposable income coefficient of residents is -23.56. is 0.08, draw a residual graph, check whether the residuals are randomly distributed around 0, and whether there is heteroscedasticity. Through analysis, it is found that the residuals are roughly randomly distributed, and there is no obvious heteroscedasticity problem;

[0058] Assuming that the model is planned to be priced at 150,000 yuan next month and the average disposable income of residents is expected to be 5,300 yuan, the trained model is used to predict the sales volume. The forecast results show that the sales volume next month is about 105 units. Based on this forecast, automobile manufacturers can reasonably arrange production plans and marketing strategies. For example, if the predicted sales volume is low, manufacturers can consider price cuts or launch new promotions to stimulate market demand.

[0059] In order to better estimate the parameters of the multiple linear regression model, the model building unit 210 estimates the parameters of the multiple linear regression model by the least square method to minimize the sum of squares of errors between actual observations and model predictions to determine the values ​​of regression coefficients;

[0060] The least squares method is a commonly used method for estimating the parameters of a multivariate linear regression model. Its core principle is to determine the regression coefficient by minimizing the sum of squares of the errors between the actual observed values ​​and the model predicted values. In the current automobile market demand forecasting case, the use of the least squares method can more accurately fit the relationship between sales quantity and sales price, and the average disposable income of residents;

[0061] Collect data containing dependent variables (such as sales volume in car sales cases) and multiple independent variables (such as sales price, average disposable income of residents). Ensure the accuracy and completeness of the data and check whether there are missing values ​​and outliers. If there are missing values, you can use mean filling, median filling or more complex interpolation methods. For outliers, you need to analyze the reasons for their generation and decide whether to correct or delete them. At the same time, according to the characteristics of the data and analysis requirements, perform pre-processing operations such as standardization and normalization on the data to make the data of different variables comparable;

[0062] The independent variables and constant terms form a design matrix. Each row represents an observation value, and each column represents a variable. According to the principle of least squares, the goal is to find the sum of squares of the errors between the predicted value and the actual observed value;

[0063] After obtaining the estimated value of the regression coefficient, the model is evaluated. Common evaluation indicators include the coefficient of determination, which measures the degree of fit of the model to the data. The value range is between 0 and 1. The closer it is to 1, the better the model fits. The adjusted coefficient of determination takes into account the impact of the number of independent variables on the fitting accuracy. The mean square error (MSE) calculates the average of the square of the error between the predicted value and the actual value. The smaller the MSE, the higher the prediction accuracy of the model. Residual analysis can also be used to check whether the residuals follow a normal distribution, whether there is autocorrelation and heteroscedasticity, etc., to judge the rationality of the model. If the model evaluation results are not ideal, it is necessary to check the data and model settings, and make adjustments and improvements.

[0064] In order to better determine the base price and the triggering conditions and adjustment range of the price, the price adjustment unit 220 uses the cost-plus pricing method to determine the base price, and determines the triggering conditions and adjustment range of the price adjustment according to the ratio between the predicted value and the actual value of the multivariate linear regression model;

[0065] The cost-plus pricing method must first comprehensively sort out the cost structure and divide the costs into fixed costs and variable costs. Fixed costs are costs that do not change with changes in production or sales, such as the purchase costs of production equipment, long-term rental costs of sites, and fixed salaries of managers. Taking automobile production as an example, the purchase costs of large-scale stamping equipment, painting production lines and other equipment used to produce automobiles, as well as the land rental costs of factories are all fixed costs. Variable costs are costs that are directly linked to production or sales, such as raw material costs and direct labor costs required to produce each product. The cost of raw materials such as steel and parts used in automobile production, as well as the wages received by workers on the production line for assembling each car, are all variable costs;

[0066] The markup rate represents the profit margin that the enterprise expects to obtain, which is affected by many factors, such as the industry average profit margin, the enterprise's own profit target, market competition conditions, etc. In the fiercely competitive automobile market, if the industry average profit margin is 20%, the enterprise may set the markup rate to 15% in order to gain market share; while some automobile enterprises with unique technology or brand advantages may increase the markup rate to 30%;

[0067] The base price is calculated based on the unit cost and the markup rate. The formula is: Base price = unit cost × (1 + markup rate). Continuing with the above example, if the markup rate is 15%, then the base price of each car is 130,000 yuan × (1 + 15%) = 149,500 yuan.

[0068] The consumption prediction module 300 includes a prediction model unit 310 and a personalized marketing unit 320;

[0069] The prediction model unit 310 divides the price adjustment data and the consumer behavior data into 70% training set and 30% test set, trains the random forest model using the training set data, adjusts the model parameters, and uses the test set to evaluate the accuracy of the model prediction;

[0070] The personalized marketing unit 320 uses the hierarchical analysis method to determine the index weights, comprehensively score the consumers, and set the stratification threshold, classifying each consumer into the corresponding level according to the comprehensive score, and generating personalized marketing suggestions;

[0071] After extracting price adjustment records and consumer behavior data from the blockchain, missing values ​​are processed first. For example, missing purchase amounts can be filled in based on the mean of consumers of the same type. Outliers, such as records with price adjustments exceeding the normal range, are corrected or deleted.

[0072] Convert consumer purchase frequency, purchase amount, and changes in purchase volume before and after price adjustment into numerical features. For example, calculate the number of purchases made by consumers in the past month as the purchase frequency feature, and use one-hot encoding for category features such as consumer gender and purchase channel. For example, if the purchase channels include online official websites, offline stores, and third-party e-commerce platforms, they will become three independent features after encoding, indicating whether the purchase was made through the channel;

[0073] Use the training set data to train the random forest model, and adjust parameters such as the number of decision trees, maximum depth, minimum number of sample splits, etc. For example, try setting the number of decision trees to 100 first, and then increase or decrease the number appropriately to optimize the model based on the performance of the model on the test set.

[0074] In order to better determine the splitting cutoff condition of the random forest model, the prediction model unit 310 measures the reduction degree of information uncertainty before and after the split node through information gain, and uses the minimum information gain threshold as the cutoff condition for the splitting of the random forest model;

[0075] Set a minimum information gain threshold. When the information gain brought by the node split is less than the threshold, the node will no longer be split. For example, if the minimum information gain is set to 0.01, if the information gain calculated by a split is less than 0.01, then the node will stop splitting to avoid the decision tree from overfitting the subtle features in the data.

[0076] In order to provide personalized marketing suggestions to consumer groups at different levels, the personalized marketing unit 320 provides personalized marketing suggestions to consumer groups at different levels, tests different marketing strategies using A / B testing, and determines the most effective marketing strategy by observing the click-through rate and conversion rate of consumers;

[0077] A / B testing is a comparative test method widely used in marketing, product design and other fields. Users are randomly divided into two groups, A and B. Solution A is applied to users in group A, and solution B is applied to users in group B. Then the behavioral data and feedback of the two groups of users are compared to determine which solution is more effective.

[0078] For different marketing strategies and content, conduct A / B testing, compare the effects of different solutions, and select the best strategy for large-scale promotion. For example, through A / B testing of different email themes, product recommendation combinations or promotional activities, observe indicators such as consumer click-through rate and conversion rate, and thus determine the most effective marketing method.

[0079] In order to optimize and update the model better, when the evaluation and optimization module 400 optimizes and updates the variables of the multivariate linear regression model using the stepwise regression method, the steps of forward selection and backward selection are alternately performed. First, the forward selection is performed to select an independent variable that makes the model improve the most from the remaining independent variables that have not been added, and then the backward selection is performed to check the independent variables that have been added to the model and eliminate the independent variable that has the smallest contribution to the model.

[0080] The forward selection method starts with an empty model that does not contain any independent variables. In consumer behavior analysis, such as studying the relationship between purchasing behavior and multiple factors, first establish a model that only contains the intercept term, and evaluate the effect of each independent variable after adding it to the model. Statistical test indicators are usually used, such as F statistics or p-values. For each independent variable, calculate the improvement in the model's goodness of fit to the data after adding it to the current model, as well as the significance of the independent variable coefficient, and select the independent variable that can maximize the improvement in the model's goodness of fit and has a significant independent variable coefficient to add to the model. For example, when considering factors such as consumer age, income, and purchase frequency, if the R² of the model is increased the most after adding the independent variable "income", and the coefficient of "income" is statistically significant, then "income" will be added to the model;

[0081] The backward selection method first includes all possible independent variables in the model. Continuing with the example of consumer behavior analysis, all relevant factors such as age, income, purchase frequency, brand preference, etc. are put into the initial model to evaluate the contribution of each independent variable to the model. Calculate the change in the goodness of fit of the model after removing each independent variable, as well as the impact on the significance of the coefficients of other independent variables, and select the independent variable that contributes the least to the model (that is, after removing the independent variable, the goodness of fit of the model decreases the least, and does not affect the significance of other independent variables) to be removed from the model. For example, in the above model, if it is found that after removing the independent variable "brand preference", the R² of the model decreases the least, and the significance of other independent variables is not affected, then "brand preference" will be removed from the model.

[0082] The second object of the present invention is to provide a blockchain-based brand marketing method, including any one of the above-mentioned blockchain-based brand marketing systems, such as Figure 3 As shown, the following steps are included:

[0083] S1. Use the market monitoring module 100 to monitor market data in real time, and use the distributed ledger to record key indicators of market supply and demand, competitive product prices, and consumer behavior, and convert them into standardized data sets;

[0084] S2. Establishing a multivariate linear regression model between market demand and price through the dynamic pricing module 200, predicting market demand under different prices, and dynamically adjusting the price of the brand with the goal of maximizing profit;

[0085] S3. Use the consumption prediction module 300 to build a random forest model to predict the consumer's future purchasing behavior and preferences, use the hierarchical analysis method to determine the indicator weights, conduct a comprehensive score on the consumer, and set a stratification threshold, classify each consumer into a corresponding level according to the comprehensive score, and generate personalized marketing suggestions;

[0086] S4. The variables of the multivariate linear regression model are optimized and updated by using the stepwise regression method through the evaluation optimization module 400, and the parameters of the random forest model are optimized and adjusted by adjusting the sample sampling ratio and feature sampling ratio of the random forest model.

[0087] In summary, the working principle of this solution is as follows:

[0088] The blockchain-based brand marketing system and method utilizes the market monitoring module 100 to monitor market data in real time, and utilizes the distributed ledger to record key indicators of market supply and demand, competitive product prices, and consumer behavior, and converts them into standardized data sets. A multivariate linear regression model between market demand and price is established through the dynamic pricing module 200 to predict market demand at different prices, and the price of the brand is dynamically adjusted with the goal of maximizing profits, so as to dynamically adjust the price of the brand according to changes in market demand, ensure that the price of the brand can keep up with market changes in real time, and achieve the purpose of brand promotion. The consumption prediction module 300 constructs a random forest model based on the price adjustment data transmitted by the dynamic pricing module 200 and the consumer behavior data transmitted by the market monitoring module 100, stratifies consumer groups, classifies each consumer into a corresponding level according to the comprehensive score, generates personalized marketing suggestions, associates the price adjustment of the brand with the consumer's consumption behavior, realizes targeted marketing to consumer groups, and improves the marketing quality and efficiency of the brand.

[0089] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. The above embodiments and descriptions are only preferred examples of the present invention and are not intended to limit the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A blockchain-based brand marketing system, characterized by: It includes a market monitoring module (100), a dynamic pricing module (200), a consumption prediction module (300) and an evaluation optimization module (400); The market monitoring module (100) monitors market data in real time through smart contracts on the blockchain platform, and uses a distributed ledger to record key indicators of market supply and demand, competitive product prices, and consumer behavior, and cleans, converts, and integrates the collected data to generate a standardized data set, and transmits the data set to the dynamic pricing module (200) and the consumption prediction module (300); The dynamic pricing module (200) establishes a multiple linear regression model between market demand and price based on the market demand data and price data recorded by the market monitoring module (100), uses the multiple linear regression model to predict market demand at different prices, and dynamically adjusts the price of the brand with the goal of maximizing profit, and transmits the price adjustment record of the brand to the consumption prediction module (300); The consumption prediction module (300) constructs a random forest model based on the price adjustment data transmitted by the dynamic pricing module (200) and the consumer behavior data transmitted by the market monitoring module (100), predicts the consumer's future purchasing behavior and preference, uses the hierarchical analysis method to determine the indicator weights, performs a comprehensive score on the consumer, sets a stratification threshold, classifies each consumer into a corresponding stratum according to the comprehensive score, and generates personalized marketing suggestions; The evaluation optimization module (400) uses the market monitoring module (100) to monitor the market data in real time, uses the stepwise regression method to optimize and update the variables of the multivariate linear regression model, and optimizes and adjusts the parameters of the random forest model by adjusting the sample sampling ratio and feature sampling ratio of the model.

2. The blockchain-based brand marketing system according to claim 1, characterized in that: The dynamic pricing module (200) includes a model building unit (210) and a price adjustment unit (220); The model building unit (210) builds a multivariate linear regression model based on the recorded market demand data and price data, calculates the mean square error between the actual observed value and the model predicted value, and uses the minimum mean square error value as the evaluation index of the model to predict the market demand under different prices; The price adjustment unit (220) calculates the profits under different price and demand combinations based on the predicted market demand and in combination with the cost data, and observes the changing trend of the profit with the price and demand by drawing the profit-price curve and the profit-demand curve, so as to dynamically adjust the price of the brand with the goal of maximizing the profit.

3. The blockchain-based brand marketing system according to claim 2, characterized in that: The formula for establishing the multivariate linear regression model in the model building unit (210) is: ； in, To meet market demand, is the intercept, For price, For seasonal factors, For price The corresponding regression coefficient is Seasonal factors The corresponding regression coefficient is is the error term.

4. The blockchain-based brand marketing system according to claim 3, characterized in that: The model building unit (210) estimates the parameters of the multivariate linear regression model by the least square method to minimize the sum of square errors between actual observations and model predicted values ​​to determine the values ​​of the regression coefficients.

5. The blockchain-based brand marketing system according to claim 2, characterized in that: The price adjustment unit (220) uses the cost-plus pricing method to determine the base price, and determines the triggering conditions for price adjustment and the price adjustment range according to the ratio between the predicted value and the actual value of the multivariate linear regression model.

6. The blockchain-based brand marketing system according to claim 1, characterized in that: The consumption prediction module (300) includes a prediction model unit (310) and a personalized marketing unit (320); The prediction model unit (310) divides the price adjustment data and the consumer behavior data into 70% training set and 30% test set, trains the random forest model using the training set data, adjusts the model parameters, and evaluates the accuracy of the model prediction using the test set; The personalized marketing unit (320) uses the hierarchical analysis method to determine the indicator weights, comprehensively score the consumers, and set the stratification threshold, so as to classify each consumer into a corresponding level according to the comprehensive score and generate personalized marketing suggestions.

7. The blockchain-based brand marketing system according to claim 6, characterized in that: The prediction model unit (310) measures the degree of reduction of information uncertainty before and after splitting a node through information gain, and uses the minimum information gain threshold as the cutoff condition for splitting the random forest model.

8. The blockchain-based brand marketing system according to claim 6, characterized in that: The personalized marketing unit (320) provides personalized marketing suggestions to consumer groups at different levels, tests different marketing strategies using A / B testing, and determines the most effective marketing strategy by observing consumers' click-through rates and conversion rates.

9. The blockchain-based brand marketing system according to claim 1, characterized in that: When the evaluation optimization module (400) optimizes and updates the variables of the multivariate linear regression model using the stepwise regression method, it alternately performs forward selection and backward selection steps, first performing forward selection to select an independent variable that improves the model the most from the remaining independent variables that have not been added, and then performing backward selection to check the independent variables that have been added to the model and eliminate the independent variable that has the smallest contribution to the model.

10. A method for implementing a blockchain-based brand marketing method, comprising a blockchain-based brand marketing system as described in any one of claims 1 to 9, characterized in that: The following steps are involved: S1. Using the market monitoring module (100) to monitor market data in real time, and using a distributed ledger to record key indicators of market supply and demand, competitive product prices, and consumer behavior, and convert them into a standardized data set; S2. Establishing a multivariate linear regression model between market demand and price through a dynamic pricing module (200), predicting market demand under different prices, and dynamically adjusting the price of the brand with the goal of maximizing profit; S3. Use the consumption prediction module (300) to build a random forest model to predict the consumer's future purchasing behavior and preference, use the hierarchical analysis method to determine the indicator weights, conduct a comprehensive score on the consumer, and set a stratification threshold to classify each consumer into a corresponding level according to the comprehensive score, and generate personalized marketing suggestions; S4. Optimizing and updating the variables of the multivariate linear regression model by using the stepwise regression method through the evaluation optimization module (400), and optimizing and adjusting the parameters of the random forest model by adjusting the sample sampling ratio and feature sampling ratio of the random forest model.

Images