Intelligent Marketing Method and System Based on User Portrait

Abstract

The present invention provides an intelligent marketing method and system based on user portraits; first, basic information and browsing behavior data of users are collected through an authorization agreement to establish an initial user portrait. Further, by comparing the basic information in the user portrait with a pre-set similarity vector, users with similar characteristics are grouped into a similar user group and share a set of similar product tags. Finally, personalized marketing information is pushed according to the similar user group where the user belongs, and at the same time, the product tag set is dynamically adjusted and optimized according to the feedback of newly launched users to ensure the continuity and effectiveness of marketing activities. This method effectively integrates user data, optimizes marketing strategies, improves the targeting and conversion rate of advertisements, and thus significantly enhances user satisfaction and the market competitiveness of enterprises.

Description

Technical Field

[0001] The present invention relates to the technical field of commercial marketing, and particularly to an intelligent marketing method and system based on user portraits. Background Art

[0002] With the rapid development of information technology, especially the application of big data, cloud computing, and artificial intelligence technologies, digital marketing has become an important core part of modern commerce. Especially in the field of e-commerce, effective marketing strategies play a crucial role in attracting and retaining users. However, traditional marketing strategies often use the same marketing information for all users, which ignores the huge differences in interests, needs, and behavior patterns among individual users, resulting in low marketing efficiency and user satisfaction.

[0003] Intelligent marketing methods attempt to provide more personalized services by constructing refined user portraits and using algorithms to predict user behavior. A user portrait is a model constructed based on multi-dimensional information such as a user's personal information, behavior data, and transaction history, which can reflect the user's preferences, needs, and potential purchase intentions. By analyzing user portraits, enterprises can identify the characteristics and needs of different user groups or individuals and achieve precision marketing.

[0004] In the prior art, intelligent marketing methods generally involve the following key steps:

[0005] [[ID=I7]]1. Data collection and processing: Collect the user's personal information, behavior data, interaction records, etc. through various channels, and these data are used to construct a user portrait after cleaning and processing.

[0006] 2. User portrait construction: Use the processed data and combine machine learning or statistical analysis methods to construct a user portrait that reflects the user's preferences and behavior patterns. This process usually includes steps such as feature selection, model training, and optimization.

[0007] 3. Marketing strategy formulation: Based on the user portrait, enterprises formulate personalized marketing strategies, which may involve product recommendations, advertising placements, promotional activities, etc.

[0008] 4. Effect evaluation and optimization: By tracking the effects of marketing activities, such as the user's click-through rate, purchase conversion rate, etc., evaluate the effectiveness of marketing strategies, and optimize the user portrait and marketing strategies according to the feedback.

[0009] Although this approach can improve the targeting and efficiency of marketing in many cases, there are still the following problems:

[0010] 1. Data fragmentation: User data is often scattered across different systems and databases, creating information silos. This not only increases the difficulty of data integration but also affects the integrity and accuracy of user profiles.

[0011] 2. Lack of real-time performance: User behaviors and preferences can change rapidly over time. In existing technologies, the update cycle of user profiles is usually long, resulting in inaccurate profile information and affecting the effectiveness of marketing campaigns.

[0012] 3. Limited personalization: Although existing technologies attempt to achieve personalized marketing through user profiles, due to the lack of in-depth behavior analysis and detailed user segmentation, this personalization is often rough and difficult to meet the personalized needs of users.

[0013] 4. Difficulty in dynamic optimization: Existing marketing systems often lack effective mechanisms to adjust marketing strategies in real time to cope with changes in the market environment and user behaviors, resulting in marketing campaigns being unable to flexibly respond to new market challenges.

[0014] To address these issues, a more efficient, dynamic, and accurate intelligent marketing method needs to be developed, which can achieve real-time data processing, dynamic user profile updates, and rapid iterative optimization of personalized marketing strategies. Summary of the Invention

[0015] To address the above issues, the object of the present invention is to propose an intelligent marketing method based on user profiles, including the following steps:

[0016] S1. Construct an initial user profile: Obtain the basic information of the user according to the authorization agreement and construct a basic information vector.

[0017] S2. Construct a single-user product label set based on the browsing behavior of a single user; Obtain the product labels involved in the user's browsing behavior according to the authorization agreement and construct a single-user product label set.

[0018] S3. Determine similar users based on the basic information vector: According to the pre-determined similarity vector, determine users with similar basic information vectors as a group of similar users.

[0019] S4. Construct a similar-user product label set based on the browsing behavior of similar users; Statistically analyze all the product labels involved in the browsing behaviors of all users within the same group of similar users and construct a similar product label set in descending order of occurrence frequency, satisfying:

[0020] ;

[0021] Wherein, represents the A set of similar product tags for similar user groups; Indicates that the product label with the highest frequency is , a total of times, and so on, The dimension of the similar product label set;

[0022] S5. Push marketing information based on the user profile of the newly-launched user and a collection of similar product tags.

[0023] Furthermore, in step S1, the basic information vector satisfies:

[0024] ;

[0025] in, Indicates the The basic information vector of each user, is the user serial number; Respectively represent The elements of the basic information vector of each user, is the vector dimension; Respectively represent Elements of a user's browsing behavior information collection, is the collection dimension.

[0026] Furthermore, in step S1, the single-user product tag set satisfies:

[0027] ;

[0028] in, Indicates the A collection of single-user product tags for each user; Respectively represent The product tags recorded by each user's browsing behavior, is the collection dimension.

[0029] Furthermore, step S3 specifically includes:

[0030] S31. Predetermine similarity vectors ,satisfy:

[0031] ;

[0032] in, Respectively represent the value ranges of the first element, the second element, and the last element of the basic information vector;

[0033] S32, all basic information vectors Falling into similar vector The users are identified as a similar user group.

[0034] Further, in step S1, the user's basic features at least include: gender, age, region, city where located, personal interest tags, and the model of the electronic device used.

[0035] Further, in step S2, the user's browsing behaviors at least include: the number of pages visited, stay time, visit frequency, click on links, click on advertisements, interaction with interface elements, search queries, click on search results, video viewing, picture browsing, picture downloading, comments, forwarding and sharing, likes, and collections.

[0036] Further, step S5 specifically includes:

[0037] S51. For newly registered users, obtain the basic information of the users according to the authorization agreement and construct a basic information vector;

[0038] S52. Determine the similar user group to which the newly registered user belongs;

[0039] S53. Retrieve the corresponding set of similar product tags for the similar user group;

[0040] S54. Push marketing information to the newly registered user according to the order of the product tags in the set of similar product tags from high to low.

[0041] Further, the intelligent marketing method based on user portraits further includes:

[0042] S6. Update the set of similar product tags according to the browsing behaviors of the newly registered user: obtain the product tags involved in the browsing behaviors of the newly registered user, supplement them to the set of similar product tags of the user group to which the newly registered user belongs, re - count the occurrence frequency of each product tag, and update the set of similar product tags.

[0043] The present invention also provides an intelligent marketing system based on user portraits for implementing the intelligent marketing method based on user portraits, including:

[0044] A user information collection module, configured to collect the basic information and browsing behaviors of users through an authorization agreement, where the basic information at least includes gender, age, region, city where located, personal interest tags, and the model of the electronic device used;

[0045] A user portrait construction module, configured to construct a user portrait according to the information collected by the user information collection module, where the user portrait includes a basic information vector and a behavior information vector;

[0046] A similar user determination module, configured to compare the user's basic information vector according to a preset similar vector to determine users with similar characteristics and group them into the same user group;

[0047] A product label set construction module, configured to construct a similar product label set according to the browsing behaviors of users within the similar user group, and the set is sorted in descending order of the frequency of occurrence of product labels;

[0048] A marketing information push module, configured to push personalized marketing information to a new user according to the user profile of the new user and the similar product label set of the affiliated similar user group;

[0049] A data update module, configured to update the corresponding similar product label set according to the browsing behaviors of new users.

[0050] The similar user determination module further includes: a vector comparison unit, configured to compare the basic information vector of a user with a preset similar vector; a user grouping unit, configured to classify the user into the corresponding user group according to the comparison result.

[0051] The beneficial effects of the present invention are as follows:

[0052] 1. Improve marketing accuracy: By constructing a detailed user profile and constructing a product label set based on the basic information and browsing behaviors of users, the specific needs and interests of users can be more accurately identified and reflected. Such a method can significantly improve the pertinence of marketing activities, thereby increasing the conversion rate.

[0053] 2. Enhance user experience: By analyzing the browsing behaviors of users and constructing similar user groups, the push of marketing information will be more in line with the actual preferences of users, reducing irrelevant advertisement interference, enhancing the user experience, and improving user satisfaction and brand loyalty.

[0054] 3. Dynamically optimize marketing strategies: By continuously monitoring the behaviors of new users and using this data to update the product label set of similar user groups, the marketing strategies can be dynamically adjusted and optimized to maintain the timeliness and relevance of marketing content.

[0055] 4. Promote personalized marketing: The system can customize personalized marketing information according to the unique profile of each user and the common characteristics of similar user groups, effectively improving user participation and interaction frequency.

[0056] 5. Optimize resource allocation: By accurately judging the similar user group to which a user belongs and pushing marketing information accordingly, marketing resources can be more effectively allocated, reducing resource waste and improving the overall ROI (return on investment) of marketing activities.

[0057] 6. Improvement of the image method: Compared with the traditional method of pushing information based on the browsing behavior of a single user, there is a problem of high randomness. The present invention calibrates similar populations through the vector method, statistically determines the common interest points of similar populations by means of big data for commodity tags, and then updates the pushed information in combination with the behavior of the user himself, improving the accuracy and fundamentality of the push. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 It is a schematic flowchart of the method of the present invention.

[0059] Figure 2 It is a running result diagram of an example program code of the method of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0060] In order to deepen the understanding of the present invention, the present invention will be further described in detail below in conjunction with embodiments. These embodiments are only used to explain the present invention and do not limit the protection scope of the present invention.

[0061] Embodiment 1

[0062] According to Figure 1 shown, this embodiment provides an intelligent marketing method based on user portraits, including the following steps:

[0063] S1. Construct an initial user portrait: Obtain the basic information of the user according to the authorization agreement and construct a basic information vector; the user's basic features at least include: gender, age, region, city where located, personal interest tags, and the model of the electronic device used; the basic information vector satisfies:

[0064] ;

[0065] Among them, represents the basic information vector of the th user, is the user serial number; respectively represent the elements of the basic information vector of the th user, is the vector dimension; respectively represent the elements of the browsing behavior information set of the th user, is the set dimension.

[0066] S2. Construct a single-user commodity tag set based on the browsing behavior of a single user; obtain the commodity tags involved in the user's browsing behavior according to the authorization agreement and construct a single-user commodity tag set;

[0067] The user browsing behavior at least includes: the number of pages visited, the stay time, the access frequency, the click on links, the click on advertisements, the interaction with interface elements, the search query, the click on search results, the video viewing, the picture browsing, the picture download, the comment, the forward sharing, the like, and the collection; the single-user commodity label set satisfies:

[0068] ;

[0069] Among them, represents the single-user commodity label set of the th user; respectively represent the commodity labels recorded by the browsing behavior of the th user, is the set dimension.

[0070] S3. Determine similar users based on the basic information vector: According to the pre-determined similar vector, the users with similar basic information vectors are determined as a group of similar users; specifically including:

[0071] S31. Pre-determine the similar vector , satisfying:

[0072] ;

[0073] Among them, respectively represent the value ranges of the first element, the second element, and until the last element of the basic information vector;

[0074] S32. Determine the users whose all basic information vectors fall into the similar vector as a group of similar users.

[0075] S4. Construct a similar user commodity label set based on the browsing behavior of similar users; count all the commodity labels involved in the browsing behavior of all users in the same group of similar users, and construct them into a similar commodity label set in the order of the highest frequency of occurrence, satisfying:

[0076] ;

[0077] Among them, represents the similar commodity label set of the th group of similar users; represents that the commodity label with the highest frequency of occurrence is , and it appears times, and so on, is the dimension of the similar commodity label set;

[0078] S5. Push marketing information based on the user portrait of newly registered users and the set of similar product tags; specifically including:

[0079] S51. For newly registered users, obtain the basic information of the users according to the authorization agreement and construct a basic information vector;

[0080] S52. Determine the similar user group to which the newly registered user belongs;

[0081] S53. Retrieve the set of corresponding similar product tags for this similar user group;

[0082] S54. Push marketing information to the newly registered user according to the order of each product tag in the set of similar product tags from high to low.

[0083] S6. Update the set of similar product tags according to the browsing behavior of newly registered users: Obtain the product tags involved in the browsing behavior of newly registered users and supplement them to the set of similar product tags of the user group to which the newly registered user belongs, re - count the occurrence frequency of each product tag, and update the set of similar product tags.

[0084] Embodiment 2

[0085] Suppose an e - commerce website wants to implement an intelligent marketing method based on user portraits. This method aims to improve the relevance and efficiency of advertisements and product recommendations, thereby increasing the user's purchase rate.

[0086] Step 1: Construct an initial user portrait

[0087] First, collect the basic information of users to construct an initial user portrait. Suppose there is a user, Mr. A, on the website, and his basic information is as follows:

[0088] Gender: Male

[0089] Age: 30 years old

[0090] City: City X

[0091] Personal interest tags: Technology, Reading, Travel

[0092] Model of electronic device used: Mobile phone model 12

[0093] This information is integrated into a basic information vector:

[0094]

[0095] Step 2: Construct a single - user product tag set based on the browsing behavior of a single user

[0096] By tracking Mr. A's website activities, we obtain the following browsing behavior data:

[0097] Number of pages visited: 15;

[0098] Stay time: 120 minutes in total;

[0099] Visit frequency: 3 times per week on average;

[0100] Number of clicks on ads: 5 times;

[0101] Number of video views: 10 times;

[0102] Number of picture views: 20 times;

[0103] Number of picture downloads: 5 times;

[0104] Number of comments: 2 times;

[0105] Number of favorites: 3 times;

[0106]

[0107] Step 3: Determine similar users based on the basic information vector:

[0108] For each type of user, we pre-define a similarity vector:

[0109] For example:

[0110] Age range: [25, 35];

[0111] Region: City X;

[0112] Step 4: Construct a set of similar user product tags based on the browsing behavior of similar users;

[0113] Mr. A is grouped into a user group that includes users of similar age ranges, the same region, and interests; by analyzing the browsing behavior of all members of the user group where Mr. A belongs, the following frequencies of product tags are found: Technology products: appear 50 times; Travel supplies: appear 30 times; Programming books: appear 20 times.

[0114] The set of similar product tags is:

[0115]

[0116] Step 5: Push marketing information:

[0117] When the new user Ms. B joins the website, based on her basic information, she is classified into the user group where Mr. A belongs. Ms. B's interests also include technology and travel. According to the set of similar product tags, ads and promotional information about technology products, travel supplies, and programming books are pushed to her in sequence.

[0118] Step 6: Update the set of similar product tags:

[0119] Ms. B's activities on the website generate new browsing data. She clicks and purchases technology products particularly frequently, so the set of similar product tags is updated to: technology products: 55 times; travel supplies: 30 times; programming books: 20 times.

[0120] This example illustrates how to achieve personalized marketing and increase user engagement and purchase intention by dynamically building and updating user profiles and product tag sets for similar user groups.

[0121] For the above embodiment, the present invention provides an exemplary program code:

[0122] import math

[0123] from collections import defaultdict

[0124] # Sample user data

[0125] users = [

[0126] {'id': 1, 'gender': 'M', 'age': 25, 'region': 'East', 'device': 'iPhone', 'browsing': ['Shoes', 'Bags', 'Hats']},

[0127] {'id': 2, 'gender': 'F', 'age': 23, 'region': 'West', 'device': 'Android', 'browsing': ['Shoes', 'Jewelry']},

[0128] {'id': 3, 'gender': 'M', 'age': 30, 'region': 'East', 'device': 'iPhone', 'browsing': ['Bags', 'Hats']},

[0129] {'id': 4, 'gender': 'F', 'age': 22, 'region': 'West', 'device': 'Android', 'browsing': ['Jewelry', 'Shoes']}, ]

[0131] # S1. Construct basic information vector

[0132] def build_basic_vector(user):

[0133] # Gender: 0 represents male, 1 represents female; Device: 0 represents iPhone, 1 represents Android

[0134] gender_value = 0 if user['gender'] == 'M' else 1

[0135] device_value = 0 if user['device'] == 'iPhone' else 1

[0136] return [gender_value, user['age'], device_value]

[0137] # S2. Build a set of product tags for a single user

[0138] def build_single_user_tags(user):

[0139] return set(user['browsing'])

[0140] # Calculate the Euclidean distance between two vectors

[0141] def calculate_distance(vec1, vec2):

[0142] return math.sqrt(sum((a - b) ** 2 for a, b in zip(vec1, vec2)))

[0143] # Snippet 3. Determine a group of similar users based on the basic information vector

[0144] def find_similar_users(users, target_user, threshold = 5):

[0145] target_vector = build_basic_vector(target_user)

[0146] similar_users = []

[0147] for user in users:

[0148] if user['id'] != target_user['id']: # Exclude the target user himself / herself

[0149] user_vector = build_basic_vector(user)

[0150] # Calculate the vector distance (Euclidean distance)

[0151] distance = calculate_distance(target_vector, user_vector)

[0152] if distance < threshold:

[0153] similar_users.append(user)

[0154] return similar_users

[0155] # S4. Build the set of product tags for similar users

[0156] def build_similar_user_tags(similar_users):

[0157] tag_counts = defaultdict(int)

[0158] for user in similar_users:

[0159] for tag in user['browsing']:

[0160] tag_counts[tag] += 1

[0161] # Sort by frequency and return the set of tags

[0162] sorted_tags = sorted(tag_counts.items(), key=lambda x: x[1], reverse=True)

[0163] return sorted_tags

[0164] # S5. Push personalized marketing information

[0165] def push_marketing_info(target_user, similar_tags):

[0166] print(f"Push the following personalized marketing information to user {target_user['id']}:")

[0167] for tag, freq in similar_tags:

[0168] print(f"Recommended product: {tag}, frequency of occurrence: {freq}")

[0169] # S51. Simulate pushing personalized marketing information for newly launched users

[0170] def marketing_for_new_user(new_user, users):

[0171] print(f"--- Build user profile, user ID: {new_user['id']} ---")

[0172] # S52. Find similar users

[0173] similar_users = find_similar_users(users, new_user)

[0174] if similar_users:

[0175] print(f"Found a group of similar users, a total of {len(similar_users)} users")

[0176] # S53. Build a set of similar product tags

[0177] similar_tags = build_similar_user_tags(similar_users)

[0178] # S54. Push personalized marketing information

[0179] push_marketing_info(new_user, similar_tags)

[0180] else:

[0181] print(f"No similar user group found.")

[0182] # Verification and use

[0183] new_user = {'id': 5, 'gender': 'F', 'age': 23,'region': 'West', 'device': 'Android', 'browsing': []}

[0184] marketing_for_new_user(new_user, users)

[0185] The verification running result of the program code is as Figure 2 shown, and the program exit code is 0, indicating that the program runs correctly without errors.

[0186] 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. What is described in the above embodiments and the specification only illustrates the principle of the present invention. Without departing from the spirit and scope of the present invention, the present invention will have various changes and improvements, and these changes and improvements all fall within the scope of the present invention claimed. The scope of protection claimed by the present invention is defined by the appended claims and their equivalents.

Claims

1. An intelligent marketing method based on user portraits, characterized in that, It includes the following steps: S1. Construct an initial user profile: Obtain the user's basic information according to the authorization agreement and construct a basic information vector; The user's basic information includes at least: gender, age, region, city where located, personal interest tags, and the model of the electronic device used; The basic information vector satisfies: ； in, Indicates the The basic information vector of each user, is the user serial number; Respectively represent The elements of the basic information vector of each user, is the vector dimension; Respectively represent Elements of a user's browsing behavior information collection, is the set dimension; The single-user product tag set satisfies: ； Among them, represents the single-user product label set of the th user; respectively represent the product labels recorded by the browsing behavior of the th user, which is the set dimension; S2. Construct a single-user product tag set based on the browsing behavior of a single user; Obtain the product tags involved in the user's browsing behavior according to the authorization agreement and construct a single-user product tag set; The user's browsing behavior includes at least: the number of pages visited, stay time, access frequency, clicked links, clicked ads, interface element interactions, search queries, search result clicks, video viewing, picture viewing, picture downloads, comments, forwarding and sharing, likes, and collections; S3. Determine similar users based on the basic information vector: According to the pre-determined similar vector, determine users with similar basic information vectors as a group of similar users; Specifically including: S31. Predetermine similar vectors , satisfying: ； Among them, respectively represent the value ranges of the first element, the second element, and up to the last element of the basic information vector; S32. Determine all basic information vectors falling into the similar vectors as a similar user group; S4. Construct a similar-user product tag set based on the browsing behavior of similar users; Count all the product tags involved in the browsing behavior of all users in the same group of similar users, and construct them into a similar product tag set in the order of the frequency of occurrence from high to low, satisfying: ； ； Among them, represents the set of similar product labels of the th similar user group; indicates that the product label with the highest frequency of occurrence is , and it appears times, and so on, is the dimension of the set of similar product labels; S5. Push marketing information according to the user profile and similar product tag set of the newly launched user.

2. The intelligent marketing method based on user portraits according to claim 1, wherein: Step S5 specifically includes: S51. For the newly launched user, obtain the user's basic information according to the authorization agreement and construct a basic information vector; S52. Determine the group of similar users to which the newly launched user belongs; S53. Retrieve the corresponding similar product tag set for this group of similar users; S54. Push marketing information to the newly launched user according to the order of each product tag in the similar product tag set from high to low.

3. The intelligent marketing method based on user portraits according to claim 2, wherein The intelligent marketing method based on the user profile further includes: S6. Update the similar product tag set according to the browsing behavior of the newly launched user: Obtain the product tags involved in the browsing behavior of the newly launched user and supplement them to the similar product tag set of the user group to which the newly launched user belongs, re-count the frequency of occurrence of each product tag, and update the similar product tag set.

4. An intelligent marketing system based on user portraits, which is used to execute an intelligent marketing method based on user portraits according to any one of claims 1-3, characterized in that, It includes: A user information collection module configured to collect the user's basic information and browsing behavior through an authorization agreement, where the basic information includes at least gender, age, region, city where located, personal interest tags, and the model of the electronic device used; A user profile construction module configured to construct a user profile according to the information collected by the user information collection module, where the user profile includes a basic information vector and a behavior information vector; A similar user determination module configured to compare the user's basic information vector according to a pre-set similar vector to determine users with similar characteristics and group them into the same user group; A product tag set construction module configured to construct a similar product tag set according to the browsing behavior of the users in the group of similar users, and this set is sorted from high to low according to the frequency of occurrence of the product tags; The marketing information push module is configured to push personalized marketing information to the user according to the user profile of the newly launched user and the set of similar product tags of the affiliated similar user group; The data update module is configured to update the corresponding set of similar product tags according to the browsing behavior of the newly launched user.

5. An intelligent marketing system based on user portraits according to claim 4, characterized in that, The similar user determination module further includes: A vector comparison unit for comparing the basic information vector of the user with a preset similar vector; A user grouping unit for classifying the user into the corresponding user group according to the comparison result.

Images