Electronic digital marketing method, system and device and storage medium

By constructing a user classification model and combining image similarity and text similarity, the problem of low efficiency and inaccurate targeting in existing digital marketing methods is solved, enabling precise marketing strategy targeting and improving marketing efficiency and accuracy.

CN118967233AInactive Publication Date: 2024-11-15HENAN IND & TRADE VOCATIONAL COLLEGE
View PDF 1 Cites 0 Cited by

Patent Information

Application Number
CN202410966928.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-18
Publication Date
2024-11-15
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

Existing digital marketing methods are inefficient and inaccurate, especially in situations where there is no ad delivery request or the request is unclear, making it difficult to effectively recommend ads or products.

Method used

By building a user classification model, training the model with historical marketing data, obtaining new user feedback data, and combining image similarity and text similarity to carry out precise marketing strategies.

Benefits of technology

It improved the accuracy and efficiency of marketing strategy deployment, ensuring precise ad placement in areas favored by users and enhancing marketing effectiveness.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118967233A_ABST
    Figure CN118967233A_ABST
Patent Text Reader

Abstract

The invention relates to the field of digital marketing, and discloses an electronic digital marketing method, system and device and a storage medium, and the method comprises the steps: obtaining historical marketing data, and marking the historical marketing data; constructing a user classification model; training the user classification model by using a historical marketing data set; for a new user, putting different advertisements to the new user in sequence, and obtaining feedback data of the new user; inputting the feedback data into the trained user classification model to obtain user classification; for different user classifications, different targeted marketing strategies are adopted. The method has the beneficial effects that before the marketing strategy is put, the users are classified based on the historical data and the feedback data to obtain the user favorite fields, so that the putting accuracy is improved; in the field of user preferences, based on a mode of fusing the image similarity and the text similarity, marketing strategies are accurately put to the users, and the marketing efficiency is improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of digital marketing, and in particular to an electronic digital marketing method, device and storage medium. Background Art

[0002] Digital marketing is a marketing strategy that promotes products and services through the internet and digital channels, with the goal of attracting target audiences, increasing awareness, and converting into sales.

[0003] Existing digital marketing methods usually use manual advertising. However, manual advertising is costly and inaccurate.

[0004] Another method uses a server to generate advertisements, such as CN114358809A, a data analysis system that applies deep neural networks to digital marketing. It uses deep neural networks to extract features from advertisement delivery requests for feature reconstruction and analyzes surface features to determine the best advertisement delivery plan. However, this method requires obtaining advertisement delivery requests first. In the absence of delivery requests or unclear delivery requests, effective advertisement or product recommendations cannot be made. Summary of the invention

[0005] The purpose of the present invention is to propose an electronic digital marketing method, device and storage medium to solve the technical problems of low efficiency and inaccurate delivery of existing digital marketing.

[0006] An electronic digital marketing method comprises the following steps:

[0007] S1: Acquire historical marketing data, and annotate the historical marketing data to obtain a marketing data set consisting of the historical marketing data;

[0008] S2: constructing a user classification model; using the historical marketing data set to train the user classification model to obtain a trained user classification model;

[0009] S3: For new users, different advertisements are served to them in sequence to obtain their feedback data;

[0010] S4: Input the feedback data into the trained user classification model to obtain user classification;

[0011] S5: Adopt different targeted marketing strategies for different user categories.

[0012] An electronic digital marketing system, comprising:

[0013] Historical data acquisition module: acquires historical marketing data, and annotates the historical marketing data to obtain a marketing data set consisting of the historical marketing data;

[0014] User classification model construction module: using the historical marketing data set to train the user classification model to obtain a trained user classification model;

[0015] Initial delivery module: for new users, different advertisements are delivered to them in sequence to obtain their feedback data;

[0016] User classification module: adopt different targeted marketing strategies for different user categories;

[0017] Marketing recommendation module: adopt different targeted marketing strategies for different user categories.

[0018] A storage medium stores instructions and data for implementing an electronic digital marketing method.

[0019] An electronic digital marketing device comprises: a processor and the storage medium; the processor loads and executes instructions and data in the storage medium to implement an electronic digital marketing method.

[0020] The beneficial effects provided by the present invention are as follows: before launching a marketing strategy, users are first classified based on historical data and feedback data to obtain the areas of user preference, thereby improving the accuracy of launch; within the areas of user preference, marketing strategies are accurately launched to users based on the fusion of image similarity and text similarity, thereby improving marketing efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 It is a schematic flow chart of the method of the present invention;

[0022] Figure 2 It is a working schematic diagram of the hardware device of the present invention. DETAILED DESCRIPTION

[0023] To make the objectives, technical solutions and advantages of the present invention more clear, the embodiments of the present invention will be further described below with reference to the accompanying drawings.

[0024] Before formally describing the present invention, the scheme of the present invention is first generally described for easy understanding.

[0025] Please refer to Figure 1 , Figure 1 It is a schematic flow diagram of the method of the present invention;

[0026] The present invention provides an electronic digital marketing method, comprising the following steps:

[0027] S1: Acquire historical marketing data, and annotate the historical marketing data to obtain a marketing data set consisting of the historical marketing data;

[0028] It should be noted that historical marketing data includes: historical user data and historical user product purchase data; product data includes product names and product categories.

[0029] It should be noted that the labeling of the historical marketing data is specifically: taking the user and the specific product as input, and labeling the product category as output.

[0030] S2: constructing a user classification model; using the historical marketing data set to train the user classification model to obtain a trained user classification model;

[0031] It should be noted that for the user classification model, a traditional decision tree, random forest, gradient boosting decision tree, logistic regression model, or a machine learning model can be used.

[0032] S3: For new users, different advertisements are served to them in sequence to obtain their feedback data;

[0033] It should be noted that for new users, since their initial status is unclear, different types of products can be presented to them in sequence to obtain feedback data such as the number of clicks, blocking status, browsing time, and types of products purchased by users.

[0034] S4: Input the feedback data into the trained user classification model to obtain user classification;

[0035] The user's feedback data includes the user's own data and data on multiple types of goods purchased by the user, or there is no product type data, only data such as the number of clicks and browsing time. Based on the feedback data and the trained user classification model, the user's classification can be obtained.

[0036] S5: Adopt different targeted marketing strategies for different user categories.

[0037] Specifically, the user classification refers to the category of goods that the user prefers to buy the most. Step S5 in the present invention is specifically as follows:

[0038] S51, for a specific user category, obtaining image features and text features of the commodities in the category;

[0039] As an example, taking sports merchandise as an example, if it is determined that the user's classification tends to be balls, the image of the ball merchandise input by the user and the corresponding text are obtained.

[0040] S52, extracting the image features of the product and calculating the image similarity;

[0041] The present invention extracts features and calculates similarities between a ball product image input by a user and other ball images.

[0042] Specifically, when performing image feature extraction, the present invention uses the VGG16 network as the corresponding extractor.

[0043] The input of the network is the user's ball product image and the comparison image whose similarity is to be calculated, and the output is the similarity value.

[0044] The feature extraction network (VGG16) uses a 1×1 convolution kernel to construct a convolution layer, and a 3×3 pooling kernel to construct a pooling layer. After the output of the last pooling layer, the vector is flattened to one dimension through a Flatten operation, and finally the feature vector is output through a fully connected layer. As an example, the calculation formula for image similarity is as follows:

[0045]

[0046] Where n is the dimension of the image feature vector; u i 、v i Represent the input image and the image to be compared respectively.

[0047] S53, extracting the text features of the product and calculating the text similarity;

[0048] As an embodiment, the present invention performs text segmentation and stop word removal on the text input by the user and the text to be compared. Furthermore, the TF-IDF method is used to count the TF-IDF values ​​of the terms; the TF-IDF values ​​are weighted averaged to complete the construction of the sentence vector;

[0049] According to the two sentence vectors, the cosine similarity is used to calculate the text similarity. The formula is as follows:

[0050] wordS=S1·S2 / (|S1|×|S2|);

[0051] S54, fusing the image similarity and the text similarity to obtain a fused similarity;

[0052] It should be noted that the fusion similarity is obtained by weighted average, specifically as follows: AS = a*wordS + b*imgS; where a+b = 1; a, b>0; a, b are preset values. In the embodiment of the present invention, according to the preliminary experimental situation, a is set to 0.6 and b is set to 0.4;

[0053] S55. Recommend corresponding marketing strategies based on fusion similarity.

[0054] In the present invention, it is set that when the fusion similarity is greater than 0.5, the picture marketing strategy is adopted, that is, the products with corresponding comparison pictures are recommended to users, and when the fusion similarity is less than 0.5, the text marketing strategy is adopted, that is, the products with corresponding texts are recommended to users.

[0055] An electronic digital marketing system, comprising:

[0056] Historical data acquisition module: acquires historical marketing data, and annotates the historical marketing data to obtain a marketing data set consisting of the historical marketing data;

[0057] User classification model construction module: using the historical marketing data set to train the user classification model to obtain a trained user classification model;

[0058] Initial delivery module: for new users, different advertisements are delivered to them in sequence to obtain their feedback data;

[0059] User classification module: adopt different targeted marketing strategies for different user categories;

[0060] Marketing recommendation module: adopt different targeted marketing strategies for different user categories.

[0061] In the historical data module, the standard process is as follows: users and specific products are taken as input, and product categories are labeled as output.

[0062] The working process of the sales recommendation module is as follows: for a specific user category, obtain the image features and text features of the products in this category; extract the image features of the products and calculate the image similarity; extract the text features of the products and calculate the text similarity; fuse the image similarity and text similarity to obtain the fused similarity; recommend the corresponding marketing strategy based on the fused similarity.

[0063] See also Figure 2 , Figure 2 It is a working diagram of the hardware device of an embodiment of the present invention, and the hardware device specifically includes: an electronic digital marketing device 401, a processor 402 and a storage medium 403.

[0064] An electronic digital marketing device 401: The electronic digital marketing device 401 implements the electronic digital marketing method.

[0065] Processor 402: The processor 402 loads and executes the instructions and data in the storage medium 403 to implement the electronic digital marketing method.

[0066] Storage medium 403: The storage medium 403 stores instructions and data; the storage medium 403 is used to implement the electronic digital marketing method.

[0067] The beneficial effects of the present invention are as follows: before launching a marketing strategy, users are first classified based on historical data and feedback data to obtain the areas of user preference, thereby improving the accuracy of the launch; within the areas of user preference, marketing strategies are accurately launched to users based on the fusion of image similarity and text similarity, thereby improving marketing efficiency.

[0068] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. An electronic digital marketing method, Features: Includes the following: S1: Acquire historical marketing data, and annotate the historical marketing data to obtain a marketing data set consisting of the historical marketing data; S2: constructing a user classification model; using the historical marketing data set to train the user classification model to obtain a trained user classification model; S3: For new users, different advertisements are served to them in sequence to obtain their feedback data; S4: Input the feedback data into the trained user classification model to obtain user classification; S5: Adopt different targeted marketing strategies for different user categories.

2. An electronic digital marketing method as claimed in claim 1, Features: The labeling of the historical marketing data specifically includes taking the user and the specific product as input and labeling the product category as output.

3. An electronic digital marketing method as claimed in claim 1, Features: Step S5 is specifically as follows: S51, for a specific user category, obtaining image features and text features of the commodities in the category; S52, extracting the image features of the product and calculating the image similarity; S53, extracting the text features of the product and calculating the text similarity; S54, fusing the image similarity and the text similarity to obtain a fused similarity; S55. Recommend corresponding marketing strategies based on fusion similarity.

4. An electronic digital marketing system, Features: include: Historical data acquisition module: acquires historical marketing data, and annotates the historical marketing data to obtain a marketing data set consisting of the historical marketing data; User classification model construction module: using the historical marketing data set to train the user classification model to obtain a trained user classification model; Initial delivery module: for new users, different advertisements are delivered to them in sequence to obtain their feedback data; User classification module: adopt different targeted marketing strategies for different user categories; Marketing recommendation module: adopt different targeted marketing strategies for different user categories.

5. An electronic digital marketing system as claimed in claim 4, Features: In the historical data module, the standard process is as follows: users and specific products are taken as input, and product categories are labeled as output.

6. An electronic digital marketing system as claimed in claim 4, Features: The working process of the marketing recommendation module is as follows: for a specific user category, obtain the image features and text features of the products in this category; extract the image features of the products and calculate the image similarity; extract the text features of the products and calculate the text similarity; fuse the image similarity and text similarity to obtain the fused similarity; recommend the corresponding marketing strategy based on the fused similarity.

7. A storage medium, Features: The storage medium stores instructions and data for implementing an electronic digital marketing method as described in any one of claims 1 to 3.

8. An electronic digital marketing device, Features: include: Processor and storage medium; the processor loads and executes instructions and data in the storage medium to implement an electronic digital marketing method as described in any one of claims 1 to 3.

Citation Information

Patent Citations

  • Data analysis system for applying deep neural network to digital marketing

    CN114358809A