Marketing information management platform and method based on big data

Abstract

The invention relates to the field of marketing information management, and particularly discloses a marketing information management platform and method based on big data, and the method comprises the steps: firstly obtaining the basic information of stationery product market clients, the purchase information of the stationery product market clients and the research data of the stationery product sales market, and then carrying out the deep learning technology, according to the method and the system, feature extraction and correlation analysis are carried out on the three, and finally, a stationery product personalized marketing strategy is generated through a generator, so that a more accurate marketing plan is made according to the characteristics and demands of different customer groups, the marketing activity is closer to the actual demands of users, the marketing efficiency is improved, and the conversion rate is increased.

Description

Technical Field

[0001] The present application relates to the field of marketing information management, and more specifically, to a marketing information management platform and method based on big data. Background Art

[0002] Stationery refers to various tools and supplies used for writing, drawing, office or study, ranging from basic pens, paper, erasers, rulers to more professional office equipment such as calculators and folders. Stationery not only plays an important role in personal learning and office environments, but is also indispensable in many fields such as education, design, and administration. Modern stationery products are diverse and their designs are constantly innovating to meet the needs of different users and enhance the user experience.

[0003] In the field of stationery marketing, given that market demand is affected by seasonal fluctuations, education cycles and economic conditions, and that different user groups such as students, teachers and office workers have significantly different specific needs for stationery products, traditional marketing methods, which usually rely on widely disseminated information to attract customers, are difficult to effectively reach and meet these diverse needs.

[0004] Therefore, a marketing information management platform and method based on big data is desired. Summary of the invention

[0005] In order to solve the above technical problems, this application is proposed. The embodiment of this application provides a marketing information management platform and method based on big data, which first obtains basic information of stationery product market customers, stationery product market customer purchase information and stationery product sales market research data, and then uses deep learning technology to perform feature extraction and correlation analysis on the three, and finally generates personalized marketing strategies for stationery products through a generator, so as to formulate more accurate marketing plans according to the characteristics and needs of different customer groups, so as to make marketing activities closer to the actual needs of users, improve marketing efficiency, and increase conversion rate.

[0006] According to one aspect of the present application, a marketing information management platform based on big data is provided, which includes:

[0007] Stationery marketing data acquisition module, used to obtain basic information of stationery product market customers, stationery product market customer purchase information and stationery product sales market research data;

[0008] A stationery marketing data extraction module is used to extract a stationery product market customer information global semantic feature vector and a stationery product sales market research associated feature vector from the stationery product market customer basic information, the stationery product market customer purchase information and the stationery product sales market research data;

[0009] A stationery marketing feature fusion module is used to fuse the global semantic feature vector of the stationery product market customer information and the stationery product sales market research related feature vector to obtain a stationery product customer marketing feature vector;

[0010] A stationery marketing feature optimization module, used for performing sparse optimization on the stationery product customer marketing feature vector based on a power series in the form of entropy degradation modulation value to obtain an optimized stationery product customer marketing feature vector;

[0011] The stationery product marketing strategy generation module is used to pass the optimized stationery product customer marketing feature vector through a generator to generate a personalized marketing strategy for the stationery product.

[0012] According to another aspect of the present application, a marketing information management method based on big data is provided, which includes:

[0013] Obtain basic information of stationery product market customers, stationery product market customer purchase information and stationery product sales market research data;

[0014] Extracting a stationery product market customer information global semantic feature vector and a stationery product sales market research associated feature vector from the stationery product market customer basic information, the stationery product market customer purchase information and the stationery product sales market research data;

[0015] Fusion of the stationery product market customer information global semantic feature vector and the stationery product sales market research associated feature vector to obtain a stationery product customer marketing feature vector;

[0016] Performing sparse optimization on the stationery product customer marketing feature vector based on the power series of entropy-degraded modulation value form to obtain an optimized stationery product customer marketing feature vector;

[0017] The optimized stationery product customer marketing feature vector is passed through a generator to generate a personalized marketing strategy for the stationery product.

[0018] Compared with the prior art, the present application provides a marketing information management platform and method based on big data, which first obtains basic customer information in the stationery product market, customer purchase information in the stationery product market, and stationery product sales market research data, and then uses deep learning technology to perform feature extraction and correlation analysis on the three, and finally uses a generator to generate personalized marketing strategies for stationery products, thereby formulating more accurate marketing plans based on the characteristics and needs of different customer groups, making marketing activities closer to users' actual needs, improving marketing efficiency, and increasing conversion rates. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] By describing the embodiments of the present application in more detail in conjunction with the accompanying drawings, the above and other purposes, features and advantages of the present application will become more apparent. The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the accompanying drawings, the same reference numerals generally represent the same components or steps.

[0020] Figure 1 It is a block diagram of a marketing information management platform based on big data according to an embodiment of the present application.

[0021] Figure 2 It is a block diagram of a stationery marketing data extraction module in a marketing information management platform based on big data according to an embodiment of the present application.

[0022] Figure 3 It is a block diagram of a stationery product market customer information feature extraction unit in a marketing information management platform based on big data according to an embodiment of the present application.

[0023] Figure 4 It is a block diagram of a stationery product sales market research feature extraction unit in a marketing information management platform based on big data according to an embodiment of the present application.

[0024] Figure 5 It is a flowchart of a marketing information management method based on big data according to an embodiment of the present application. DETAILED DESCRIPTION

[0025] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described here.

[0026] Figure 1 Schematic diagram of a marketing information management platform based on big data according to an embodiment of the present application. Figure 1As shown, according to the marketing information management platform 100 based on big data of the embodiment of the present application, it includes: a stationery marketing data acquisition module 110, which is used to obtain basic information of stationery product market customers, purchase information of stationery product market customers and stationery product sales market research data; a stationery marketing data extraction module 120, which is used to extract the global semantic feature vector of stationery product market customer information and the associated feature vector of stationery product sales market research from the basic information of stationery product market customers, the purchase information of stationery product market customers and the stationery product sales market research data; a stationery marketing feature fusion module 130, which is used to fuse the global semantic feature vector of stationery product market customer information and the associated feature vector of stationery product sales market research to obtain a stationery product customer marketing feature vector; a stationery marketing feature optimization module 140, which is used to perform sparse optimization on the stationery product customer marketing feature vector based on the power series in the form of entropy degradation modulation value to obtain an optimized stationery product customer marketing feature vector; a stationery product marketing strategy generation module 150, which is used to pass the optimized stationery product customer marketing feature vector through a generator to generate a personalized marketing strategy for stationery products.

[0027] In the above-mentioned marketing information management platform 100 based on big data, the stationery marketing data acquisition module 110 is used to obtain basic information of customers in the stationery product market, customer purchase information in the stationery product market, and stationery product sales market research data. It should be understood that stationery refers to various tools and supplies used for writing, painting, office or learning, covering from basic pens, paper, erasers, rulers to more professional office equipment such as calculators and folders. There are many types of modern stationery products, and the design is constantly innovating to meet the needs of different users and improve the user experience. In the field of stationery marketing, given that market demand is affected by seasonal fluctuations, education cycles and economic conditions, and different user groups such as students, teachers and office staff have significant differences in specific needs for stationery products, traditional marketing methods usually rely on widely disseminated information to attract customers, and it is difficult to effectively reach and meet these diverse needs. That is, the demand for stationery in the market is affected by many factors, including seasonal fluctuations, education cycles and economic conditions. For example, the start of school and exam seasons usually bring a significant increase in demand for stationery, while it is relatively stable at other times. In addition, the demand for stationery at different educational stages (such as primary school, middle school, and university) is also different, requiring targeted marketing strategies. Economic conditions directly affect consumers' purchasing power, and users of different economic levels have different price sensitivities to stationery. Different user groups such as students, teachers, and office workers have different needs and preferences for stationery, requiring personalized products and services. Students may pay more attention to price and practicality, while office workers may pay more attention to product quality and brand. In addition, traditional marketing methods usually rely on widely disseminated information such as advertising and promotions, making it difficult to accurately reach target users, and the needs of different user groups vary significantly, making it difficult for traditional marketing methods to provide personalized solutions.

[0028] Specifically, obtaining basic information about customers in the stationery product market is the basis for understanding the characteristics of customer groups. This information includes the customer's age, gender, occupation, geographic location, etc. By analyzing this basic information, companies can identify the characteristics and needs of different customer groups and develop targeted marketing strategies. For example, students may pay more attention to price and practicality, while office workers may pay more attention to product quality and brand. Customer purchase information in the stationery product market includes customer purchase history, purchase frequency, and purchase amount. By analyzing purchase history, companies can identify customer loyalty and purchase cycle, and develop corresponding customer maintenance strategies. Purchase frequency and purchase amount can help companies evaluate customers' spending power and potential value. Stationery product sales market research data includes market trends, competitor analysis, product demand forecasts, etc. Market trend analysis can help companies understand the current market hot spots and potential opportunities, so as to adjust product lines and marketing strategies. Competitor analysis can help companies identify competitive advantages and disadvantages and develop differentiated competitive strategies. Product demand forecasts can help companies plan production and inventory in advance to avoid imbalances in supply and demand. By comprehensively acquiring basic information on stationery product market customers, stationery product market customer purchase information, and stationery product sales market research data, companies can build a comprehensive market analysis framework to generate highly personalized marketing strategies.

[0029] In the above-mentioned marketing information management platform 100 based on big data, the stationery marketing data extraction module 120 is used to extract the stationery product market customer information global semantic feature vector and the stationery product sales market research related feature vector from the stationery product market customer basic information, the stationery product market customer purchase information and the stationery product sales market research data. In this way, the enterprise can conduct in-depth mining and analysis on these feature vectors, thereby generating highly personalized marketing strategies to help enterprises optimize operational efficiency and enhance market competitiveness.

[0030] Figure 2 FIG. 1 is a block diagram of a stationery marketing data extraction module in a marketing information management platform based on big data according to an embodiment of the present application. Figure 2 As shown, in a specific embodiment of the present application, the stationery marketing data extraction module 120 includes: a stationery product market customer information feature extraction unit 121, used to perform feature extraction on the stationery product market customer basic information and the stationery product market customer purchase information to obtain the stationery product market customer information global semantic feature vector; a stationery product sales market research feature extraction unit 122, used to perform feature extraction on the stationery product sales market research data to obtain the stationery product sales market research associated feature vector.

[0031] It should be understood that the basic customer information in the stationery product market includes static attributes such as the customer's age, gender, occupation, and geographic location. By segmenting and encoding this basic information, companies can obtain semantic feature vectors of the customer's basic information. This process involves natural language processing (NLP) technology and deep learning models such as bidirectional long short-term memory neural network (BiLSTM). In this way, companies can capture the potential semantic features in the customer's basic information, such as the impact of the customer's age group, occupation type, and geographic location on the demand for stationery.

[0032] Furthermore, stationery product sales market research data usually exists in text form, including macro market information such as market trends, competitor analysis, and product demand forecasts. These data are unstructured and difficult to use directly for data analysis and machine learning models. Through feature extraction, companies can convert stationery product sales market research data into stationery product sales market research related feature vectors. This process converts unstructured text data into structured data, which is convenient for subsequent feature extraction and analysis.

[0033] Figure 3 FIG. 1 is a block diagram of a stationery product market customer information feature extraction unit in a marketing information management platform based on big data according to an embodiment of the present application. Figure 3 As shown, in a specific embodiment of the present application, the stationery product market customer information feature extraction unit 121 includes: a customer basic information feature encoding subunit 1211, used to feature encode the stationery product market customer basic information to obtain a stationery product market customer basic information semantic feature vector; a customer purchase information feature encoding subunit 1212, used to feature encode the stationery product market customer purchase information to obtain a stationery product market customer purchase information semantic feature vector; a customer information global feature association subunit 1213, used to associate the stationery product market customer basic information semantic feature vector and the stationery product market customer purchase information semantic feature vector to obtain the stationery product market customer information global semantic feature vector.

[0034] It should be understood that encoding the basic information of customers in the stationery product market to obtain the semantic feature vector of the basic information of customers in the stationery product market can help enterprises deeply understand the static attributes of customer groups and identify the characteristics and needs of different customer groups. For example, students may pay more attention to price and practicality, while office workers may pay more attention to product quality and brand.

[0035] Furthermore, feature encoding of customer purchase information in the stationery product market is performed to obtain semantic feature vectors of customer purchase information in the stationery product market, which can help companies deeply understand customers' dynamic behaviors and consumption habits, and also provide a scientific basis for formulating personalized marketing strategies. Among them, based on semantic feature vectors, companies can formulate personalized marketing strategies and accurately push products and services that meet customer behavior patterns. For example, for customers with high frequency of purchases, companies can design membership discounts and point rewards; for customers with low frequency of purchases, companies can design high-quality and professional product recommendations.

[0036] Furthermore, the semantic feature vector of basic customer information in the stationery product market is obtained by segmenting and encoding basic customer information (such as age, gender, occupation, geographic location, etc.). These feature vectors reflect the static attributes of customers and help companies understand their potential needs and preferences. The semantic feature vector of customer purchase information in the stationery product market is obtained by contextual encoding and word-granular feature extraction of customer purchase information (such as purchase history, purchase frequency, purchase amount, etc.). By associating the semantic feature vector of basic customer information with the semantic feature vector of purchase information, companies can obtain a global semantic feature vector of customer information in the stationery product market. This global feature vector combines the static attributes and dynamic behaviors of customers, providing companies with comprehensive and in-depth customer insights.

[0037] In a specific embodiment of the present application, the customer basic information feature encoding subunit 1211 includes: segmenting the stationery product market customer basic information to obtain a text sequence of the stationery product market customer basic information; passing the stationery product market customer basic information text sequence through a stationery product market customer basic information bidirectional long short-term memory neural network model to obtain a semantic feature vector of the stationery product market customer basic information.

[0038] It should be understood that word segmentation of basic customer information in the stationery product market helps enterprises convert unstructured text data into structured data, providing a scientific basis for formulating personalized marketing strategies. Specifically, basic customer information in the stationery product market usually exists in text form, including static attributes such as customer age, gender, occupation, and geographic location. This information is unstructured and difficult to be directly used for data analysis and machine learning models. For example, customer occupation information may be recorded in the form of free text, such as "teacher", "student", "engineer", etc. These text data cannot be directly input into machine learning models for processing. Through word segmentation, enterprises can decompose the customer basic information text into independent words or phrases to form a text sequence of basic customer information in the stationery product market. The text sequence after word segmentation can be input into a natural language processing (NLP) model for feature extraction. For example, through the bidirectional long short-term memory neural network (Bi LSTM) model, enterprises can capture the potential semantic features in the customer basic information, such as the impact of the customer's age group, occupation type, and geographic location on the demand for stationery.

[0039] Furthermore, the bidirectional long short-term memory neural network model (Bi LSTM) is a deep learning model that can capture long-distance dependencies in sequence data and is suitable for processing text data. Among them, the Bi LSTM model can capture the potential semantic features in the text sequence and help enterprises deeply understand the semantic information of the basic information of customers. For example, through the Bi LSTM model, enterprises can identify the impact of customers' age, occupation type and geographical location on the demand for stationery. In addition, the BiLSTM model can capture long-distance dependencies in the text sequence and help enterprises understand the complex relationships in the basic information of customers. For example, the occupation type and geographical location of customers may have a mutual impact on the demand for stationery, and the Bi LSTM model can capture these relationships. By inputting the text sequence of basic information of customers in the stationery product market into the Bi LSTM model, the enterprise can obtain the semantic feature vector of the basic information of customers in the stationery product market. .

[0040] In a specific embodiment of the present application, the customer purchase information feature encoding subunit 1212 includes: passing the stationery product market customer purchase information through a stationery product market customer purchase information context encoder to obtain multiple stationery product market customer purchase information text feature vectors; arranging the multiple stationery product market customer purchase information text feature vectors in one dimension to obtain a stationery product market customer purchase information one-dimensional text feature vector; passing the stationery product market customer purchase information one-dimensional text feature vector through a stationery product market customer purchase information one-dimensional text word granularity feature extractor to obtain the stationery product market customer purchase information semantic feature vector.

[0041] It should be understood that the context encoder of customer purchase information in the stationery product market is a model that can capture contextual relationships in sequence data and is suitable for processing purchase behavior data. By inputting the semantic feature vector of the basic information of customers in the stationery product market into the context encoder, the enterprise can obtain multiple text feature vectors of customer purchase information in the stationery product market. The context encoder can capture the contextual relationship between the semantic feature vector of the basic information of customers and the purchase behavior data, and help enterprises deeply understand the purchase behavior and consumption habits of customers. Specifically, the customer purchase information in the stationery product market is segmented to obtain a customer purchase information word sequence; the stationery product market customer purchase information stationery product sales market research text semantic encoder is used to map each customer purchase information word in the customer purchase information word sequence to a word embedding vector to obtain a sequence of customer purchase information word embedding vectors; the converter-based Bert model of the stationery product market customer purchase information context encoder is used to perform global context semantic encoding on the sequence of customer purchase information word embedding vectors to obtain multiple stationery product market customer purchase information text feature vectors.

[0042] Furthermore, multiple text feature vectors of customer purchase information in the stationery product market are obtained by contextual encoding of customer purchase information (such as purchase history, purchase frequency, purchase amount, etc.). These feature vectors are usually multi-dimensional and difficult to use directly for subsequent feature extraction and analysis. For example, a customer's purchase history may contain purchase records at multiple time points, which form a multi-dimensional text feature vector. Through one-dimensional arrangement, the enterprise can convert multiple text feature vectors of customer purchase information in the stationery product market into one-dimensional text feature vectors of customer purchase information in the stationery product market. One-dimensional arrangement converts multi-dimensional text feature vectors into one-dimensional feature vectors, reduces the dimension of the features, and facilitates subsequent feature extraction and analysis. For example, purchase records at multiple time points are converted into one-dimensional purchase frequency feature vectors.

[0043] Furthermore, the one-dimensional text feature vector of customer purchase information in the stationery product market is passed through a one-dimensional text word granularity feature extractor for customer purchase information in the stationery product market to help enterprises deeply understand customers' purchasing behavior and consumption habits. Among them, the one-dimensional text word granularity feature extractor is a model that can capture local features in sequence data and is suitable for processing one-dimensional text feature vectors. By inputting the one-dimensional text feature vector of customer purchase information in the stationery product market into the one-dimensional text word granularity feature extractor, enterprises can obtain semantic feature vectors of customer purchase information in the stationery product market. The one-dimensional text word granularity feature extractor can capture local features in one-dimensional text feature vectors, helping enterprises deeply understand customers' purchasing behavior and consumption habits. For example, through a one-dimensional convolutional neural network model, enterprises can capture local features such as customer purchase frequency, purchase amount, and purchase cycle. Specifically, each layer of the one-dimensional text word granularity feature extractor of the stationery product market customer purchase information is used to perform convolution processing, mean pooling processing based on the local feature matrix and non-linear activation processing on the input data in the forward pass of the layer, so that the last layer of the one-dimensional text word granularity feature extractor of the stationery product market customer purchase information outputs the semantic feature vector of the stationery product market customer purchase information, wherein the input of the one-dimensional text word granularity feature extractor of the stationery product market customer purchase information is the one-dimensional text feature vector of the stationery product market customer purchase information.

[0044] In a specific embodiment of the present application, the customer information global feature association subunit 123 includes: creating an item calculate_global_feature_vector; representing the stationery product market customer basic information semantic feature vector and the stationery product market customer purchase information semantic feature vector as basic_info_vector and purchase_info_vector; performing a transposition operation on the stationery product market customer purchase information semantic feature vector purchase_info_vector to generate a transposed vector purchase_info_vector_transposed of the stationery product market customer purchase information semantic feature vector; calculating the transposed vector purchase_info_vector_transposed of the stationery product market customer basic information semantic feature vector basic_info_vector and the stationery product market customer purchase information semantic feature vector purchase_info_vector_transposed; transposed product to obtain the global semantic feature vector global_feature_vector of customer information in the stationery product market; encapsulate the association operation into a SemanticFeatureVectorService service, and deploy the SemanticFeatureVectorService service to the Eureka server for discovery and calling; use the Flask framework to implement the SemanticFeatureVectorService service, define the route calculate_global_feature_vector to receive customer basic information and purchase information, and return the global semantic feature vector global_feature_vector of customer information in the stationery product market; run the Flask application on port 5000, and use the py_eureka_client library to register the service to the Eureka server.

[0045] It should be understood that in this way, a global semantic feature vector calculation service for customer information in the stationery product market based on deep learning is implemented and deployed to the Eureka server. The main function of this service is to convert the customer's basic information and purchase information into a semantic feature vector, and obtain the global semantic feature vector through matrix multiplication.

[0046] The py_eureka_client library is used to initialize the Eureka client and deploy the service registration to the Eureka server. The service name is SemanticFeatureVectorService and the running port is 5000.

[0047] Use the transformers library to load the pre-trained BERT model and tokenizer. The BERT model is used to convert text data into semantic feature vectors, and the tokenizer is used to perform tokenization on the input text.

[0048] Use the Flask framework to implement the Web service and define the route calculate_global_feature_vector to receive the basic customer information and purchase information and return the global semantic feature vector. In the route processing function, first obtain the basic customer information and purchase information, then use the BERT model to generate the semantic feature vector. Then transpose the purchase information semantic feature vector, and finally calculate the global semantic feature vector and return it.

[0049] Define an auxiliary function generate_semantic_vector to convert the input text into a semantic feature vector. This function uses the BERT model to process the input text and returns the feature vector after average pooling.

[0050] The numpy library is used to perform matrix multiplication operations to calculate the product of the semantic feature vector of the customer's basic information and the transposed vector of the semantic feature vector of the purchase information to obtain the global semantic feature vector.

[0051] In this way, an efficient and scalable global semantic feature vector calculation service for customer information in the stationery product market is realized. Among them, the BERT model is used to extract semantic features of customer basic information and purchase information, which can capture the deep semantic information in the text and generate high-quality semantic feature vectors. The global semantic feature vector is obtained by calculating the product of the semantic feature vector of customer basic information and the transposed vector of the semantic feature vector of purchase information through matrix multiplication. This process can comprehensively consider the basic information and purchase behavior of customers and generate a comprehensive and integrated feature vector, which is helpful for subsequent market analysis and personalized recommendations. The service is registered to the Eureka server using the Eureka mechanism, which realizes automatic discovery and calling of the service. This makes the service easy to expand and deploy, and improves the maintainability and scalability of the system. The Flask framework is used to implement the Web service interface, which provides a simple and easy-to-use API for client calls. The client only needs to send JSON data containing the basic information and purchase information of the customer to obtain the global semantic feature vector. The numpy library is used for matrix multiplication operations, which has high computational efficiency and can quickly process large-scale data. At the same time, the pre-training characteristics of the BERT model make the feature extraction process efficient and accurate.

[0052] Specifically, part of the deployment code is shown below.

[0053]

[0054] # Initialize BERT model and word segmenter

[0055] tokenizer=BertTokenizer.from_pretrained('bert-base-uncased')

[0056] model=BertModel.from_pretrained('bert-base-uncased')

[0057] app = Flask(__name__)

[0058] defgenerate_semantic_vector(text):

[0059] inputs=tokenizer(text,return_tensors='pt',truncation=True,padding=True)

[0060] outputs = model(**inputs)

[0061] return outputs.last_hidden_state.mean(dim=1).detach().numpy()

[0062] @app.route(' / calculate_global_feature_vector',methods=['POST'])

[0063] defcalculate_global_feature_vector():

[0064] data=request.json

[0065] #Get basic customer information and purchase information

[0066] basic_info=data['basic_info']

[0067] purchase_info=data['purchase_info']

[0068] #Generate semantic feature vector

[0069] basic_info_vector=generate_semantic_vector(basic_info)

[0070] purchase_info_vector=generate_semantic_vector(purchase_info)

[0071] #Transpose the semantic feature vector of purchase information

[0072] purchase_info_vector_transposed=purchase_info_vector.T

[0073] #Calculate the global semantic feature vector

[0074] global_feature_vector=np.dot(basic_info_vector, purchase_info_vector_transposed)

[0075]

[0076] More specifically, this realizes a powerful and easily scalable global semantic feature vector calculation service for customer information in the stationery product market, which can effectively improve the accuracy and efficiency of market analysis and personalized recommendations.

[0077] Figure 4 FIG. 1 is a block diagram of a stationery product sales market research feature extraction unit in a marketing information management platform based on big data according to an embodiment of the present application. Figure 4 As shown, in a specific embodiment of the present application, the stationery product sales market research feature extraction unit 122 includes: a stationery product sales market research semantic encoding subunit 1221, used to pass the stationery product sales market research data through a stationery product sales market research text semantic encoder to obtain a plurality of stationery product sales market research semantic feature vectors; a stationery market research feature splicing subunit 1222, used to splice the plurality of stationery product sales market research semantic feature vectors to obtain the stationery product sales market research associated feature vector.

[0078] It should be understood that stationery product sales market research data usually exists in text form, including macro market information such as market trends, competitor analysis, and product demand forecasts. These data are unstructured and difficult to use directly for data analysis and machine learning models. For example, market trends may be recorded in the form of free text, and these text data cannot be directly input into machine learning models for processing. The text semantic encoder is a model that can capture semantic information in text data and is suitable for processing market research data. By inputting stationery product sales market research data into the text semantic encoder, enterprises can obtain multiple stationery product sales market research semantic feature vectors. Among them, the text semantic encoder can capture the potential semantic features in the market research data, helping enterprises to deeply understand market dynamics and trends. Based on multiple semantic feature vectors, enterprises can obtain comprehensive market insights and identify market opportunities and potential risks. Specifically, the stationery product sales market survey data is converted into an embedding vector using the stationery product sales market survey text semantic encoder to obtain a sequence of stationery product sales market survey embedding vectors; and the sequence of stationery product sales market survey embedding vectors is subjected to global contextual semantic encoding using the Bert model of the stationery product sales market survey text semantic encoder to obtain multiple stationery product sales market survey semantic feature vectors.

[0079] Furthermore, multiple stationery product sales market research semantic feature vectors are obtained by text semantic encoding of market research data (such as market trends, competitor analysis, product demand forecasts, etc.). Splicing is a key step in integrating multi-dimensional semantic feature vectors into an associated feature vector. Through splicing, an enterprise can convert multiple stationery product sales market research semantic feature vectors into stationery product sales market research associated feature vectors. Splicing integrates multi-dimensional semantic feature vectors into an associated feature vector, helping enterprises to comprehensively consider multiple dimensions in market research data. For example, features such as market trends, competitor analysis, and product demand forecasts are integrated into an associated feature vector. In a specific embodiment of the present application, the generator model is a variational autoencoder (VAE).

[0080] In the above-mentioned marketing information management platform 100 based on big data, the stationery marketing feature fusion module 130 is used to fuse the global semantic feature vector of the stationery product market customer information and the stationery product sales market research associated feature vector to obtain the stationery product customer marketing feature vector. It should be understood that the global semantic feature vector of the stationery product market customer information is obtained by extracting and associating the basic information and purchase information of the customer. These feature vectors reflect the static attributes and dynamic behaviors of the customer, helping the enterprise to understand the potential needs and preferences of the customer. The stationery product sales market research associated feature vector is obtained by text semantic encoding and splicing of the market research data. These feature vectors reflect the macro environment and future trends of the market, helping enterprises to understand market dynamics and trends. Fusion is a key step to integrate the global semantic feature vector of the stationery product market customer information and the stationery product sales market research associated feature vector into a customer marketing feature vector. Through fusion, enterprises can comprehensively consider customer needs and market dynamics to obtain stationery product customer marketing feature vectors. The fused customer marketing feature vector captures the correlation between customer needs and market dynamics, helping enterprises to deeply understand the complex relationship between customer needs and market trends. For example, the correlation between customers' occupation types and market trends can help companies predict customers' future needs and market opportunities.

[0081] In the above-mentioned marketing information management platform 100 based on big data, the stationery marketing feature optimization module 140 is used to perform sparse optimization on the stationery product customer marketing feature vector based on the power series in the form of entropy degenerate modulation value to obtain an optimized stationery product customer marketing feature vector.

[0082] In particular, in the technical solution of the present application, when processing the basic information and purchase information of customers in the stationery product market, the two data sets may contain repeated features. For example, the customer's purchasing habits may overlap with the age, gender and other features in their basic information, which generates redundant information. This redundancy not only increases the dimension of the data, but also leads to an increase in the correlation between features, thereby reducing the independence of the features. Further, the processing of sales market research data may also introduce redundant features. The survey data may contain content related to the customer's basic information and purchasing behavior, such as customer preferences, popular trends, etc., which may be repeated with the features in the first two data sources during the feature fusion process. Therefore, the increase in redundant features leads to the densification of the feature space, making it difficult for the generator to capture effective patterns when generating samples. Considering that the generator relies on effective feature combinations to generate personalized marketing strategies, redundant features may cause the model to focus too much on certain specific feature combinations and ignore other potentially important features. This situation makes the generated samples tend to cluster in the feature space, lacking the necessary diversity, resulting in a reduced effectiveness of the generated results. In this way, the generated personalized marketing strategy cannot accurately reflect the real needs of different customers. This not only reduces the iteration effect of the generator, but also may make the final marketing strategy lack of pertinence, reducing the market's response efficiency. Therefore, in the technical solution of the present application, the stationery product customer marketing feature vector is sparsely optimized based on the power series of entropy-degraded modulation value form to obtain an optimized stationery product customer marketing feature vector.

[0083] Among them, the stationery product customer marketing feature vector is subjected to sparse optimization based on a power series in the form of entropy degradation modulation value to obtain an optimized stationery product customer marketing feature vector, including: extracting each row vector in the prediction parameter space of the generator to obtain a set of stationery product customer marketing prediction parameter row vectors; calculating the stationery product customer marketing probability entropy degradation simulation value between the stationery product customer marketing feature vector and each stationery product customer marketing prediction parameter row vector in the set of stationery product customer marketing prediction parameter row vectors to obtain a set of stationery product customer marketing probability entropy degradation simulation values; calculating the square root of the sum of squares of the stationery product customer marketing probability entropy degradation simulation values ​​to obtain a stationery product customer marketing entropy degradation modulation value; based on the stationery product customer marketing entropy degradation modulation value, performing feature distribution modulation on the stationery product customer marketing feature vector to obtain the optimized stationery product customer marketing feature vector.

[0084] The optimization formula is:

[0085] D j =Softmax(M m v c +H m pj +b m )

[0086]

[0087] Among them, P represents the set of row vectors of stationery product customer marketing prediction parameters, p1, p2, p j 、p n represents the first, second, jth, and nth stationery product customer marketing prediction parameter row vectors in the set of stationery product customer marketing prediction parameter row vectors, v c represents the customer marketing feature vector of the stationery product, Softmax represents the normalized exponential function, M m represents the first weight matrix, H m represents the second weight matrix, b m represents the bias vector, D j represents the jth stationery product customer marketing probability entropy degradation simulation value of the set of stationery product customer marketing probability entropy degradation simulation values, t represents the stationery product customer marketing entropy degradation modulation value, v c ⊙t represents the calculation of the customer marketing feature vector v of stationery products c Each eigenvalue of the t-th power, represents the i-th eigenvalue of the customer marketing feature vector of the stationery product, α represents a predetermined hyperparameter, represents the addition by position, and V' represents the optimized customer marketing feature vector of stationery products.

[0088] Specifically, in the technical solution of the present application, the stationery product customer marketing feature vector has feature redundancy, which causes a problem of sample squeeze relative to the generator when it is generated by the generator, thus affecting the sample iteration effect.

[0089] Based on this, in the technical solution of the present application, the entropy degradation of the stationery product customer marketing feature vector based on the prediction parameter space of the generator is maximized. First, each row vector is extracted from the prediction parameter space of the generator to form a row vector set. Then, the stationery product customer marketing probability entropy degradation simulation value between the stationery product customer marketing feature vector and each row vector in the above row vector set is calculated to measure the uncertainty or information entropy change of the stationery product customer marketing feature vector under given parameters. Furthermore, all the stationery product customer marketing probability entropy degradation simulation values ​​are squared and summed, and then the square root is taken to obtain a comprehensive stationery product customer marketing entropy degradation modulation value, which reflects the degree of entropy change of the entire system when processing the feature vector. Finally, the stationery product customer marketing feature vector is adjusted according to the stationery product customer marketing entropy degradation modulation value obtained above. Specifically, the attribute set of the stationery product customer marketing feature vector is represented by a formal power series based on the stationery product customer marketing entropy degradation modulation value, and a set aggregation bounded deviation represented by an aggregation radius based on each attribute value of the stationery product customer marketing feature vector is added, so as to perform attribute selection and sparse optimization on the stationery product customer marketing feature vector under a unified stationery product customer marketing entropy degradation modulation value restriction standard, so as to ensure that the stationery product customer marketing feature vector contains sufficient data but is not so cumbersome as to introduce unnecessary interference or redundancy.

[0090] In the above-mentioned marketing information management platform 100 based on big data, the stationery product marketing strategy generation module 150 is used to pass the optimized stationery product customer marketing feature vector through a generator to generate a personalized marketing strategy for stationery products. It should be understood that the optimized stationery product customer marketing feature vector is obtained by integrating the global semantic feature vector of the stationery product market customer information and the associated feature vector of the stationery product sales market research. These feature vectors integrate the static attributes and dynamic behaviors of customers, as well as the macro environment and future trends of the market, to help companies fully understand customer needs and market dynamics. Among them, the generator is a model that can generate personalized marketing strategies based on input feature vectors, and is suitable for processing optimized stationery product customer marketing feature vectors. Specifically, by inputting the optimized stationery product customer marketing feature vector into the generator, the company can generate a personalized marketing strategy for stationery products, thereby helping the company to accurately push products and services that meet customer needs and market trends.

[0091] In summary, the embodiment of the present application first obtains basic information of stationery product market customers, stationery product market customer purchase information and stationery product sales market research data, and then uses deep learning technology to perform feature extraction and correlation analysis on the three, and finally uses a generator to generate personalized marketing strategies for stationery products, so as to formulate more accurate marketing plans based on the characteristics and needs of different customer groups, so that marketing activities are closer to the actual needs of users, improve marketing efficiency, and increase conversion rates.

[0092] As described above, the marketing information management platform 100 based on big data according to the embodiment of the present application can be implemented in various terminal devices. In one example, the marketing information management platform 100 based on big data can be integrated into the terminal device as a software module and / or a hardware module. For example, the marketing information management platform 100 based on big data can be a software module in the operating system of the terminal device, or can be an application developed for the terminal device; of course, the marketing information management platform 100 based on big data can also be one of the many hardware modules of the terminal device.

[0093] Alternatively, in another example, the big data-based marketing information management platform 100 and the terminal device may also be separate devices, and the big data-based marketing information management platform 100 may be connected to the terminal device via a wired and / or wireless network and transmit interactive information in accordance with an agreed data format.

[0094] Figure 5 FIG. 1 is a flowchart of a marketing information management method based on big data according to an embodiment of the present application. Figure 5 As shown, according to the marketing information management method based on big data of the embodiment of the present application, it includes: S110, obtaining basic information of stationery product market customers, purchase information of stationery product market customers and stationery product sales market research data; S120, extracting the global semantic feature vector of stationery product market customer information and the associated feature vector of stationery product sales market research from the basic information of stationery product market customers, the purchase information of stationery product market customers and the stationery product sales market research data; S130, fusing the global semantic feature vector of stationery product market customer information and the associated feature vector of stationery product sales market research to obtain a stationery product customer marketing feature vector; S140, performing sparse optimization on the stationery product customer marketing feature vector based on the power series in the form of entropy degenerate modulation value to obtain an optimized stationery product customer marketing feature vector; S150, passing the optimized stationery product customer marketing feature vector through a generator to generate a personalized marketing strategy for stationery products.

[0095] Here, those skilled in the art can understand that the specific operations of each step in the above-mentioned marketing information management method based on big data have been referred to above. Figures 1 to 4 The description of the marketing information management platform based on big data has been introduced in detail, and therefore, its repeated description will be omitted.

[0096] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0097] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0098] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0099] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0100] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0101] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A marketing information management platform based on big data, characterized in that: include: Stationery marketing data acquisition module, used to obtain basic information of stationery product market customers, stationery product market customer purchase information and stationery product sales market research data; A stationery marketing data extraction module is used to extract a stationery product market customer information global semantic feature vector and a stationery product sales market research associated feature vector from the stationery product market customer basic information, the stationery product market customer purchase information and the stationery product sales market research data; A stationery marketing feature fusion module is used to fuse the global semantic feature vector of the stationery product market customer information and the stationery product sales market research related feature vector to obtain a stationery product customer marketing feature vector; A stationery marketing feature optimization module, used for performing sparse optimization on the stationery product customer marketing feature vector based on a power series in the form of entropy degradation modulation value to obtain an optimized stationery product customer marketing feature vector; The stationery product marketing strategy generation module is used to pass the optimized stationery product customer marketing feature vector through a generator to generate a personalized marketing strategy for the stationery product.

2. The marketing information management platform based on big data according to claim 1, characterized in that: The stationery marketing data extraction module includes: A stationery product market customer information feature extraction unit, used for extracting features from the stationery product market customer basic information and the stationery product market customer purchase information to obtain a global semantic feature vector of the stationery product market customer information; The stationery product sales market research feature extraction unit is used to extract features from the stationery product sales market research data to obtain the stationery product sales market research related feature vector.

3. The marketing information management platform based on big data according to claim 2, characterized in that: The stationery product market customer information feature extraction unit comprises: A customer basic information feature encoding subunit, used for performing feature encoding on the basic information of customers in the stationery product market to obtain a semantic feature vector of the basic information of customers in the stationery product market; A customer purchase information feature encoding subunit, used for performing feature encoding on the customer purchase information of the stationery product market to obtain a semantic feature vector of the customer purchase information of the stationery product market; The customer information global feature association subunit is used to associate the semantic feature vector of the basic information of the stationery product market customers with the semantic feature vector of the purchase information of the stationery product market customers to obtain the global semantic feature vector of the stationery product market customer information.

4. The marketing information management platform based on big data according to claim 3 is characterized in that: The customer basic information feature encoding subunit includes: Segmenting the basic information of customers in the stationery product market to obtain a text sequence of the basic information of customers in the stationery product market; The text sequence of the basic information of customers in the stationery product market is passed through a bidirectional long short-term memory neural network model of the basic information of customers in the stationery product market to obtain a semantic feature vector of the basic information of customers in the stationery product market.

5. The marketing information management platform based on big data according to claim 4 is characterized in that: The customer purchase information feature encoding subunit includes: Passing the stationery product market customer purchase information through a stationery product market customer purchase information context encoder to obtain a plurality of stationery product market customer purchase information text feature vectors; One-dimensionally arranging the plurality of stationery product market customer purchase information text feature vectors to obtain a one-dimensional text feature vector of the stationery product market customer purchase information; The one-dimensional text feature vector of the stationery product market customer purchase information is passed through a one-dimensional text word granularity feature extractor of the stationery product market customer purchase information to obtain a semantic feature vector of the stationery product market customer purchase information.

6. The marketing information management platform based on big data according to claim 5, characterized in that: The customer information global feature association subunit includes: Create project calculate_global_feature_vector; The semantic feature vector of the basic information of the stationery product market customers and the semantic feature vector of the purchase information of the stationery product market customers are represented as basic_info_vector and purchase_info_vector; Performing a transposition operation on the semantic feature vector purchase_info_vector of the customer purchase information in the stationery product market to generate a transposed vector purchase_info_vector_transposed of the semantic feature vector of the customer purchase information in the stationery product market; Calculate the product of the semantic feature vector of basic information of the stationery product market customer basic_info_vector and the transposed vector purchase_info_vector_transposed of the semantic feature vector of the stationery product market customer purchase information to obtain the global semantic feature vector of the stationery product market customer information global_feature_vector; Encapsulate the association operation as a SemanticFeatureVectorService service, and deploy the SemanticFeatureVectorService service to the Eureka server for discovery and invocation; The SemanticFeatureVectorService service is implemented using the Flask framework, and the route calculate_global_feature_vector is defined to receive basic customer information and purchase information, and return the global semantic feature vector global_feature_vector of the customer information in the stationery product market; Run the Flask application on port 5000 and register the service with the Eureka server using the py_eureka_client library.

7. The marketing information management platform based on big data according to claim 6, characterized in that: The stationery product sales market research feature extraction unit comprises: A stationery product sales market research semantic encoding subunit, used for passing the stationery product sales market research data through a stationery product sales market research text semantic encoder to obtain a plurality of stationery product sales market research semantic feature vectors; The stationery market research feature splicing subunit is used to splice the multiple stationery product sales market research semantic feature vectors to obtain the stationery product sales market research related feature vector.

8. The marketing information management platform based on big data according to claim 7, characterized in that: The stationery marketing feature optimization module includes: Extract each row vector in the prediction parameter space of the generator to obtain a set of row vectors of stationery product customer marketing prediction parameters; Calculating the stationery product customer marketing probability entropy degradation simulation value between the stationery product customer marketing feature vector and each stationery product customer marketing prediction parameter row vector in the set of stationery product customer marketing prediction parameter row vectors to obtain a set of stationery product customer marketing probability entropy degradation simulation values; Calculating the square root of the sum of squares of the stationery product customer marketing probability entropy degradation simulation values ​​to obtain the stationery product customer marketing entropy degradation modulation value; Based on the stationery product customer marketing entropy degradation modulation value, the stationery product customer marketing feature vector is subjected to feature distribution modulation to obtain the optimized stationery product customer marketing feature vector.

9. A marketing information management method based on big data, characterized in that: include: Obtain basic information of stationery product market customers, stationery product market customer purchase information and stationery product sales market research data; Extracting a stationery product market customer information global semantic feature vector and a stationery product sales market research associated feature vector from the stationery product market customer basic information, the stationery product market customer purchase information and the stationery product sales market research data; Fusion of the stationery product market customer information global semantic feature vector and the stationery product sales market research associated feature vector to obtain a stationery product customer marketing feature vector; Performing sparse optimization on the stationery product customer marketing feature vector based on the entropy-degraded modulation value form power series to obtain an optimized stationery product customer marketing feature vector; The optimized stationery product customer marketing feature vector is passed through a generator to generate a personalized marketing strategy for the stationery product.

10. The marketing information management method based on big data according to claim 9, characterized in that: Extracting a stationery product market customer information global semantic feature vector and a stationery product sales market research associated feature vector from the stationery product market customer basic information, the stationery product market customer purchase information and the stationery product sales market research data, including: Performing feature extraction on the basic information of the stationery product market customers and the purchase information of the stationery product market customers to obtain a global semantic feature vector of the stationery product market customer information; Feature extraction is performed on the stationery product sales market research data to obtain the stationery product sales market research related feature vector.

Images