Medical intelligent marketing system based on SCRM

Abstract

The invention, which belongs to the technical field of marketing in the medical industry, discloses an SCRM-based medical intelligent marketing system comprising a data acquisition module, a data processing and analysis module, an AI recommendation engine module, a client interaction platform module and an effect evaluation and feedback mechanism module. According to the method, an official API provided by each social media platform is utilized to guide a user to access public information of the user through an OAuth2.0 authorization protocol and an authorization application, then an HTTP request library is used to call the API, the public information of the user is acquired, and sensitive data is encrypted by using a symmetric encryption technology; real-time data such as search records and page browsing records of a user are collected through a front-end application and a website, when it is detected that the user searches for related information of a certain disease, a recommendation system is triggered to generate new recommendation content, and then according to search keywords of the user, the new recommendation content is recommended. Matching related popular science articles from the content library, and pushing the matched content to the user through a popup window and a notification bar after the expert lectures the video; and multi-language switching is supported in the front-end application, so that the user can select different language versions to meet the requirements of international patients.

Description

Technical Field

[0001] The present invention belongs to the field of marketing technology in the medical industry, and specifically refers to a medical intelligent marketing system based on SCRM. Background Art

[0002] With the acceleration of digital transformation, the service model of the medical industry is undergoing a profound transformation from the traditional model to digital and intelligent. On the one hand, patients' demands for medical services are becoming more diversified and personalized; on the other hand, medical institutions are facing challenges such as how to effectively reach target groups, improve patient stickiness, and optimize resource allocation. In this context, it is particularly important to develop a medical intelligent marketing system that can integrate online and offline resources, deeply analyze user behavior, and provide personalized services.

[0003] However, the existing technology is insufficient in data integration capabilities. Existing medical information systems are often scattered across different platforms and departments, with serious data silos. As a result, data cannot be effectively integrated and shared, and data formats and standards from different sources are not unified, which increases the difficulty of data integration and affects data quality and availability. Existing technologies mainly rely on users' static information and a small amount of behavioral data, lacking multi-dimensional dynamic data support, resulting in user portraits that are not comprehensive and accurate enough. At the same time, in the process of constructing user portraits, there are insufficient measures to protect user privacy, which can easily cause users to worry about data security. For this reason, a medical intelligent marketing system based on SCRM is proposed. Summary of the invention

[0004] The purpose of the present invention is to provide a medical intelligent marketing system based on SCRM to solve the problems raised in the above background technology.

[0005] To achieve the above-mentioned purpose, the present invention provides the following technical solutions: a medical intelligent marketing system based on SCRM, including a data acquisition module, a data processing and analysis module, an AI recommendation engine module, a customer interaction platform module, and an effect evaluation and feedback mechanism module;

[0006] The data collection module can collect multi-dimensional data on patients’ personal information, health status, and consumption behavior from multiple channels including hospital information systems, health applications, and social media;

[0007] The data processing and analysis module pre-processes the collected data, performs data cleaning, format conversion, and outlier detection operations, and uses advanced data analysis technology and machine learning algorithms to deeply mine the data and build detailed user portraits;

[0008] The AI ​​recommendation engine uses deep learning technology to provide personalized medical services or product recommendations based on user portraits and historical interaction data. The recommended content covers disease prevention, health management, and disease treatment.

[0009] The customer interaction platform supports multiple communication methods, including instant messaging, telephone and email, to promote effective communication between medical staff and patients;

[0010] The effect evaluation and feedback mechanism quantifies the effect of marketing activities by setting a KPI indicator system, and continuously optimizes the recommendation model and service process based on user feedback.

[0011] Among them, the data collection module uses the official API interface provided by major social media platforms to guide users through the OAuth 2.0 authorization protocol, authorize applications to access their public information, and then use the HTTP request library to call the API interface to obtain the user's public information; at the same time, it is compatible with third-party health application interfaces, and is compatible with the relevant data information transmitted and obtained by the interfaces of third-party applications such as wearable devices and health management apps, so as to obtain more comprehensive patient health data through the same authorized access method; when verifying data, a hash function is used to calculate the data summary, compare whether the summaries of the two data sources are consistent, and use symmetric encryption technology to encrypt sensitive data. The encryption formula is divided into four steps. Step 1: initial round key addition, the implementation formula is:

[0012]

[0013] In the formula, P is the plaintext block (128 bits), K0 is the initial key (128 bits), and C0 is the output of the initial round; Step 2: Multiple rounds of transformation, the implementation formula is:

[0014] S i,j =S box [S i,j ]

[0015] In the formula, S i,j is the byte in row i and column j of the state matrix, S box is the S-box table; Step 3: Final round transformation, the implementation formula is:

[0016]

[0017] In the formula, W[i] is the i-th word, N k The word number T(W[i-1]) of the key is a transformation function, including byte replacement, circular shift and round constant XOR. The symmetric encryption algorithm is used to achieve efficient encryption and decryption of data. These transformations include byte replacement, row shift, column mixing and key addition steps, ensuring the high security and anti-attack capability of the algorithm.

[0018] The data processing and analysis module uses natural language processing and machine learning algorithms to parse unstructured data, uses the TF-IDF method to extract text features and then selects the support vector machine SVM machine learning algorithm for training, and the implementation formula is:

[0019] TF-IDF(t,d0=TF(t,d)×IDF(t)

[0020] In the formula, TF(t,d) is the frequency of word t appearing in document d, and IDF(t) is the inverse document frequency, calculated as:

[0021]

[0022] In the formula, N is the total number of documents, n t is the number of documents containing word t; the trained model is used to identify the potential needs of patients and possible health risks, which are valuable features for marketing decisions; at the same time, sentiment analysis technology is introduced to select the naive Bayes model for model training through machine learning methods, and the implementation formula is:

[0023]

[0024] In the formula, P(C k |x) is the category C under the given feature vector x k The posterior probability, P(x|C k ) is Category C k The conditional probability of the eigenvector x under k ) is Category C k The prior probability of P(x) is the marginal probability of the feature vector x. The trained model is used to classify the sentiment of the new text, judge the positive, negative and neutral sentiment, evaluate the patient's attitude towards specific services and products, and assist in the formulation of marketing strategies. The graph neural network technology is used to build a relationship network between users, and the GCN model is used to learn the embedded representation of the node. The implementation formula is:

[0025]

[0026] In the formula, H (l+1) is the node feature matrix of the lth layer, is the adjacency matrix with self-loops added, yes The degree matrix, W (l)is the weight matrix of the lth layer, σ is the activation function, which discovers the potential community effect and provides support for group marketing; these technical means and implementation steps together constitute an efficient and intelligent data processing and analysis system, providing strong support for other modules of the medical intelligent marketing system.

[0027] Among them, the AI ​​recommendation engine module can dynamically adjust the recommended content according to the real-time needs of users. First, the real-time data such as user search records and page browsing records are collected through the front-end application and the website. When it is detected that the user searches for relevant information about a certain disease, the recommendation system is triggered to generate new recommended content. Then, according to the user's search keywords, relevant popular science articles and expert lecture videos are matched from the content library, and the matched content is pushed to the user through pop-up windows and notification bars. The formula for calculating the similarity matching between keywords and content libraries is:

[0028]

[0029] In the formula, A and B are two vectors, representing the TF-IDF vectors of two documents respectively, and · represents the

[0030] Quantify dot product, ∥A∥

[0031] and ∥B∥ represent the modulus of vectors A and B respectively; obtain the user's geographic location information through the user's IP address and GPS positioning technology, update the user portrait based on the user's geographic location and time factors, add geographic and time dimension features, and generate recommendations based on the user's historical behavior and the behavior of similar users through the collaborative filtering formula. The implementation formula is:

[0032]

[0033] In the formula, is the predicted rating of user u for item i, is the average rating of user u, sin(u,v) is the similarity between user u and user v, r vi is the actual rating of item i by user v, is the average rating of user v, N(u,i) is the set of users similar to user u who have rated item i, and the generated recommended content is pushed to the user; the recommendation engine detects the user's browser settings and the language preference set by the user through NLP technology, and uses the machine translation model Transformer to translate the recommended content into the language selected by the user. The implementation formula is:

[0034]

[0035] In the formula, Q is the query vector, K is the key vector, V is the value vector, and d kis the dimension of the key vector, softmax is the softmax function, which is used to normalize the probability distribution. Multi-language switching is supported in the front-end application, and users can choose different language versions to meet the needs of international patients.

[0036] The customer interaction platform module has intelligent customer service functions, collects and organizes common questions and their answers, builds a knowledge base, uses support vector machines (SVMs) for intent recognition, and identifies user intent. The implementation formula is:

[0037]

[0038] In the formula, y is the user's intention, x is the user's input text, Y is the set of all possible intentions, and P(y|x) is the probability of intention y given the input text x. At the same time, the module supports automatic speech recognition to realize the recognition and understanding of speech input. The front end converts the user's speech input into text, cleans and pre-processes the converted text, uses the speech recognition formula to identify the user's intention and extract key entities from the text, and finally selects the appropriate answer and performs the corresponding operation based on the user's intention and entity. The implementation formula is:

[0039]

[0040] In the formula, is the recognized text, X is the input speech signal, P(w|X) is the probability of text w given the speech signal X; the implementation formula for entity extraction is:

[0041]

[0042] In the formula, y i is the entity tag of the i-th word, x i is the i-th word, x i-1 and x i+1 are adjacent words, Y is the set of all possible entity tags, P(y i |x i ,x i-1 ,x i+1 ) is the probability of an entity label given a word and its context; through intelligent customer service supporting speech recognition and natural language understanding technology, it can accurately capture the user's intention and make appropriate responses.

[0043] Among them, the effect evaluation and feedback mechanism module defines a series of KPI indicators according to the marketing goals, including click-through rate, conversion rate, and follow-up rate. First, the user's click count, purchase count, and follow-up count data are collected through the front-end application, and the collected data is cleaned, deduplicated, and format converted to ensure data quality. Then, the values ​​of various indicators are calculated according to the defined KPI formula. Finally, the calculation results are displayed in the form of charts and dashboards for easy viewing and analysis by the management and marketing team. At the same time, the effects of different marketing strategies are compared through A / B testing to find the best solution. The module randomly divides users into multiple groups, each group uses different marketing strategies, and runs marketing activities of multiple groups at the same time in the same environment, records the user behavior data of each group, uses statistical methods to analyze the data of each group, evaluates the effects of different strategies, and selects the best performing strategy as the final solution based on the data analysis results. The chi-square test is used to compare whether there are significant differences in the classification variables of multiple groups. The implementation formula is:

[0044]

[0045] In the formula, O i is the observed value, E i is the expected value, n is the number of categorical variables; the last module will regularly collect and summarize the data of marketing activities, conduct in-depth analysis of the summarized data, identify the reasons for success and failure, and write detailed marketing reports, including data charts, analysis conclusions, and recommended measures. The marketing report will be issued to relevant management and marketing teams through emails and meetings, and feedback from relevant personnel will be collected. The report content and analysis methods will be continuously improved, lessons learned will be summarized, and future marketing work will be guided.

[0046] The system comprises the following steps:

[0047] S1. Obtain patient-related information through the data collection module to ensure that the data source is legal and compliant;

[0048] S2. Generate user portraits using data processing and analysis modules, including basic information, health status, and consumption habits;

[0049] S3. AI recommendation engines provide personalized services or product recommendations based on user portraits. Recommendations must be strictly reviewed to ensure their scientificity and legality.

[0050] S4. Realize doctor-patient communication through the customer interaction platform and provide convenient consulting services;

[0051] S5. Use the effect evaluation and feedback mechanism to monitor the effectiveness of marketing activities, adjust strategies in a timely manner, and improve marketing efficiency.

[0052] Among them, in step S1: the data collection process strictly abides by the relevant laws and regulations of the "Cybersecurity Law" and the "Personal Information Protection Law" to ensure information security and personal privacy protection; and clearly informs users of the purpose, scope and usage rules of data collection, and respects users' right to know and right to choose.

[0053] Among them, in step S3: personalized recommendations not only take into account user preferences, but also take into account the professionalism and safety of medical services to avoid damaging patient interests due to excessive marketing; for special groups (such as minors and the elderly), recommendations will be more cautious to prevent misleading or adverse effects; and medical experts are regularly invited to review the recommendations to ensure that they comply with the latest medical standards and ethical norms.

[0054] Among them, in step S5: in addition to the conventional marketing effect evaluation, it also includes a survey on patient satisfaction, which serves as an important reference for optimizing services; conducting regular user surveys to collect feedback on system functions, interface design, ease of operation, etc., and continuously improving user experience; establishing a sound complaint handling mechanism, giving positive responses to users' reasonable demands, and maintaining a good brand image.

[0055] Compared with the prior art, the present invention has the following beneficial effects:

[0056] 1. The present invention can integrate multi-source data from different platforms and departments through the data acquisition module, break the data silos, realize comprehensive data sharing and efficient use, and ensure the consistency and high quality of data by establishing unified data standards and formats, thereby improving the availability and reliability of data. At the same time, the system supports real-time data stream processing, which can capture the latest behaviors and needs of users in a timely manner, and ensure the timeliness and effectiveness of marketing strategies;

[0057] 2. The present invention not only considers the user's static information (such as age and gender), but also integrates the user's dynamic behavior data (such as search history and medical records) to build a comprehensive and accurate user portrait. It uses machine learning and natural language processing data mining technology to deeply analyze the user's real needs and potential preferences, thereby improving the accuracy and satisfaction of personalized services;

[0058] 3. The present invention combines deep learning, graph neural network and other recommendation algorithms to improve the diversity and accuracy of recommendations. By introducing content-based recommendations and new user behavior analysis, it effectively solves the cold start problem of new users or new items and improves the accuracy of initial recommendations. In addition, the system can dynamically adjust the recommended content according to the user's real-time behavior to ensure that the recommended content always meets the user's current needs and interests.

[0059] 4. The present invention has established a KPI indicator system covering multiple aspects such as click-through rate, conversion rate, and re-visit rate, which comprehensively and accurately measures the performance of marketing activities. The system supports scientific A / B testing, and by comparing the effects of different marketing strategies, the optimal solution is found, which improves the scientificity and effectiveness of the marketing strategy. At the same time, the system can also quickly collect and process user feedback, adjust the marketing strategy in a timely manner, and improve the system's response speed and market adaptability;

[0060] 5. The present invention focuses on the friendliness of the user interface and interactive design, provides a simple and intuitive operating interface, improves the user experience and satisfaction, and the system supports multiple language versions, which can meet the needs of international patients and expand the application scope and user group of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] Figure 1 The operation process of the medical intelligent marketing system based on SCRM of the present invention Figure 1 ;

[0062] Figure 2 The operation process of the medical intelligent marketing system based on SCRM of the present invention Figure 2 ;

[0063] Figure 3 The operation process of the medical intelligent marketing system based on SCRM of the present invention Figure 3 . DETAILED DESCRIPTION

[0064] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0065] For examples, see Figure 1-Figure 3 As shown, the present invention provides a technical solution: including a data acquisition module, a data processing and analysis module, an AI recommendation engine module, a customer interaction platform module, and an effect evaluation and feedback mechanism module;

[0066] The data collection module can collect multi-dimensional data on patients’ personal information, health status, and consumption behavior from multiple channels including hospital information systems, health applications, and social media;

[0067] The data processing and analysis module pre-processes the collected data, performs data cleaning, format conversion, and outlier detection operations, and uses advanced data analysis technology and machine learning algorithms to deeply mine the data and build detailed user portraits;

[0068] The AI ​​recommendation engine uses deep learning technology to provide personalized medical services or product recommendations based on user portraits and historical interaction data. The recommended content covers disease prevention, health management, and disease treatment.

[0069] The customer interaction platform supports multiple communication methods, including instant messaging, telephone and email, to promote effective communication between medical staff and patients;

[0070] The effect evaluation and feedback mechanism quantifies the effect of marketing activities by setting a KPI indicator system, and continuously optimizes the recommendation model and service process based on user feedback.

[0071] Among them, the data collection module uses the official API interface provided by major social media platforms to guide users through the OAuth 2.0 authorization protocol, authorize applications to access their public information, and then use the HTTP request library to call the API interface to obtain the user's public information; at the same time, it is compatible with third-party health application interfaces, and is compatible with the relevant data information transmitted and obtained by the interfaces of third-party applications such as wearable devices and health management apps, so as to obtain more comprehensive patient health data through the same authorized access method; when verifying data, a hash function is used to calculate the data summary, compare whether the summaries of the two data sources are consistent, and use symmetric encryption technology to encrypt sensitive data. The encryption formula is divided into four steps. Step 1: initial round key addition, the implementation formula is:

[0072]

[0073] In the formula, P is the plaintext block (128 bits), K0 is the initial key (128 bits), and C0 is the output of the initial round; Step 2: Multiple rounds of transformation, the implementation formula is:

[0074] S i,j =S box [S i,j ]

[0075] In the formula, S i,j is the byte in row i and column j of the state matrix, S box is the S-box table; Step 3: Final round transformation, the implementation formula is:

[0076]

[0077] In the formula, W[i] is the i-th word, N kThe word number T(W[i-1]) of the key is a transformation function, including byte replacement, circular shift and round constant XOR. The symmetric encryption algorithm is used to achieve efficient encryption and decryption of data. These transformations include byte replacement, row shift, column mixing and key addition steps, ensuring the high security and anti-attack capability of the algorithm.

[0078] The data processing and analysis module uses natural language processing and machine learning algorithms to parse unstructured data, uses the TF-IDF method to extract text features and then selects the support vector machine SVM machine learning algorithm for training, and the implementation formula is:

[0079] TF-IDF(t,d)=TF(t,d)×IDF(t)

[0080] In the formula, TF(t,d) is the frequency of word t appearing in document d, and IDF(t) is the inverse document frequency, calculated as:

[0081]

[0082] In the formula, N is the total number of documents, n t is the number of documents containing word t; the trained model is used to identify the potential needs of patients and possible health risks, which are valuable features for marketing decisions; at the same time, sentiment analysis technology is introduced to select the naive Bayes model for model training through machine learning methods, and the implementation formula is:

[0083]

[0084] In the formula, P(C k |x) is the category C under the given feature vector x k The posterior probability, P(x|C k ) is Category C k The conditional probability of the eigenvector x under k ) is Category C k The prior probability of P(x) is the marginal probability of the feature vector x. The trained model is used to classify the sentiment of the new text, judge the positive, negative and neutral sentiment, evaluate the patient's attitude towards specific services and products, and assist in the formulation of marketing strategies. The graph neural network technology is used to build a relationship network between users, and the GCN model is used to learn the embedded representation of the node. The implementation formula is:

[0085]

[0086] In the formula, H (l+1) is the node feature matrix of the lth layer, is the adjacency matrix with self-loops added, yes The degree matrix, W (l) is the weight matrix of the lth layer, σ is the activation function, which discovers the potential community effect and provides support for group marketing; these technical means and implementation steps together constitute an efficient and intelligent data processing and analysis system, providing strong support for other modules of the medical intelligent marketing system.

[0087] Among them, the AI ​​recommendation engine module can dynamically adjust the recommended content according to the real-time needs of users. First, the real-time data such as user search records and page browsing records are collected through the front-end application and the website. When it is detected that the user searches for relevant information about a certain disease, the recommendation system is triggered to generate new recommended content. Then, according to the user's search keywords, relevant popular science articles and expert lecture videos are matched from the content library, and the matched content is pushed to the user through pop-up windows and notification bars. The formula for calculating the similarity matching between keywords and content libraries is:

[0088]

[0089] In the formula, A and B are two vectors, representing the TF-IDF vectors of two documents respectively, and · represents the

[0090] Quantify dot product, ∥A∥

[0091] and ∥B∥ represent the modulus of vectors A and B respectively; obtain the user's geographic location information through the user's IP address and GPS positioning technology, update the user portrait based on the user's geographic location and time factors, add geographic and time dimension features, and generate recommendations based on the user's historical behavior and the behavior of similar users through the collaborative filtering formula. The implementation formula is:

[0092]

[0093] In the formula, is the predicted rating of user u for item i, is the average rating of user u, sin(u,v) is the similarity between user u and user v, r vi is the actual rating of item i by user v, is the average rating of user v, N(u,i) is the set of users similar to user u who have rated item i, and the generated recommended content is pushed to the user; the recommendation engine detects the user's browser settings and the language preference set by the user through NLP technology, and uses the machine translation model Transformer to translate the recommended content into the language selected by the user. The implementation formula is:

[0094]

[0095] In the formula, Q is the query vector, K is the key vector, V is the value vector, and dk is the dimension of the key vector, softmax is the softmax function, which is used to normalize the probability distribution. Multi-language switching is supported in the front-end application, and users can choose different language versions to meet the needs of international patients.

[0096] The customer interaction platform module has intelligent customer service functions, collects and organizes common questions and their answers, builds a knowledge base, uses support vector machines (SVMs) for intent recognition, and identifies user intent. The implementation formula is:

[0097]

[0098] In the formula, y is the user's intention, x is the user's input text, Y is the set of all possible intentions, and P(y|x) is the probability of intention y given the input text x. At the same time, the module supports automatic speech recognition to realize the recognition and understanding of speech input. The front end converts the user's speech input into text, cleans and pre-processes the converted text, uses the speech recognition formula to identify the user's intention and extract key entities from the text, and finally selects the appropriate answer and performs the corresponding operation based on the user's intention and entity. The implementation formula is:

[0099]

[0100] In the formula, is the recognized text, X is the input speech signal, P(w|X) is the probability of text w given the speech signal X; the implementation formula for entity extraction is:

[0101]

[0102] In the formula, y i is the entity tag of the i-th word, x i is the i-th word, x i-1 and x i+1 are adjacent words, Y is the set of all possible entity tags, P(y i |x i ,x i-1 ,x i+1 ) is the probability of an entity label given a word and its context; through intelligent customer service supporting speech recognition and natural language understanding technology, it can accurately capture the user's intention and make appropriate responses.

[0103] Among them, the effect evaluation and feedback mechanism module defines a series of KPI indicators according to the marketing goals, including click-through rate, conversion rate, and follow-up rate. First, the user's click count, purchase count, and follow-up count data are collected through the front-end application, and the collected data is cleaned, deduplicated, and format converted to ensure data quality. Then, the values ​​of various indicators are calculated according to the defined KPI formula. Finally, the calculation results are displayed in the form of charts and dashboards for easy viewing and analysis by the management and marketing team. At the same time, the effects of different marketing strategies are compared through A / B testing to find the best solution. The module randomly divides users into multiple groups, each group uses different marketing strategies, and runs marketing activities of multiple groups at the same time in the same environment, records the user behavior data of each group, uses statistical methods to analyze the data of each group, evaluates the effects of different strategies, and selects the best performing strategy as the final solution based on the data analysis results. The chi-square test is used to compare whether there are significant differences in the classification variables of multiple groups. The implementation formula is:

[0104]

[0105] In the formula, O i is the observed value, E i is the expected value, n is the number of categorical variables; the last module will regularly collect and summarize the data of marketing activities, conduct in-depth analysis of the summarized data, identify the reasons for success and failure, and write detailed marketing reports, including data charts, analysis conclusions, and recommended measures. The marketing report will be issued to relevant management and marketing teams through emails and meetings, and feedback from relevant personnel will be collected. The report content and analysis methods will be continuously improved, lessons learned will be summarized, and future marketing work will be guided.

[0106] The system comprises the following steps:

[0107] S1. Obtain patient-related information through the data collection module to ensure that the data source is legal and compliant;

[0108] S2. Generate user portraits using data processing and analysis modules, including basic information, health status, and consumption habits;

[0109] S3. AI recommendation engines provide personalized services or product recommendations based on user portraits. Recommendations must be strictly reviewed to ensure their scientificity and legality.

[0110] S4. Realize doctor-patient communication through the customer interaction platform and provide convenient consulting services;

[0111] S5. Use the effect evaluation and feedback mechanism to monitor the effectiveness of marketing activities, adjust strategies in a timely manner, and improve marketing efficiency.

[0112] Among them, in step S1: the data collection process strictly abides by the relevant laws and regulations of the "Cybersecurity Law" and the "Personal Information Protection Law" to ensure information security and personal privacy protection; and clearly informs users of the purpose, scope and usage rules of data collection, and respects users' right to know and right to choose.

[0113] Among them, in step S3: personalized recommendations not only take into account user preferences, but also take into account the professionalism and safety of medical services to avoid damaging patient interests due to excessive marketing; for special groups (such as minors and the elderly), recommendations will be more cautious to prevent misleading or adverse effects; and medical experts are regularly invited to review the recommendations to ensure that they comply with the latest medical standards and ethical norms.

[0114] Among them, in step S5: in addition to the conventional marketing effect evaluation, it also includes a survey on patient satisfaction, which serves as an important reference for optimizing services; conducting regular user surveys to collect feedback on system functions, interface design, ease of operation, etc., and continuously improving user experience; establishing a sound complaint handling mechanism, giving positive responses to users' reasonable demands, and maintaining a good brand image.

[0115] Working principle: The system uses the official API interfaces provided by major social media platforms to obtain users' public information through authorized access, guides users to authorize applications to access their public information through the OAuth 2.0 authorization protocol, uses the HTTP request library to call the API interface, obtains users' public information, parses the returned JSON or XML format data into usable structured data, and finally stores the parsed data in the MongoDB database;

[0116] The user's search history and page browsing history are collected through front-end applications and websites in real time. When it is detected that the user is searching for information related to a certain disease, the recommendation system is triggered to generate new recommended content. According to the user's search keywords, relevant popular science articles and expert lecture videos are matched from the content library, and the matched content is pushed to the user. The system obtains the user's geographic location information through the user's IP address and GPS positioning technology, records the user's behavior time, such as access time and search time, and updates the user portrait according to the user's geographic location and time factors, adding features of geographic and time dimensions, and generates personalized recommended content in combination with the user's geographic location and time factors.

[0117] The customer interaction platform module is equipped with natural language processing and dialogue management system technologies to realize the function of automatically answering common questions. The system collects and organizes common questions and their answers, builds a knowledge base, uses natural language understanding technology to identify user intentions, selects appropriate answers or performs corresponding operations based on user intentions and context, generates natural language replies, and sends them to users in the form of text or voice;

[0118] The system will define a series of KPI indicators such as click-through rate, conversion rate, and repeat visit rate based on marketing goals. It will collect users' behavioral data on clicks, purchases, and repeat visits through front-end applications, and perform pre-processing operations such as cleaning, deduplication, and format conversion on the collected data to ensure data quality. It will calculate the values ​​of various indicators based on the defined KPI formulas, and display the calculation results in the form of charts, dashboards, etc., for easy viewing and analysis by management and the marketing team.

[0119] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

[0120] The present invention and its embodiments are described above, and such description is not restrictive. The drawings show only one embodiment of the present invention, and the actual structure is not limited thereto. In short, if ordinary technicians in the field are inspired by it, without departing from the purpose of the invention, they can design a structure and embodiment similar to the technical solution without creativity, which should belong to the protection scope of the present invention.

Claims

1. The medical intelligent marketing system based on SCRM includes data collection module, data processing and analysis module, AI recommendation engine module, customer interaction platform module and effect evaluation and feedback mechanism module; The data collection module can collect multi-dimensional data on patients’ personal information, health status, and consumption behavior from multiple channels including hospital information systems, health applications, and social media; The data processing and analysis module pre-processes the collected data, performs data cleaning, format conversion, and outlier detection operations, and uses advanced data analysis technology and machine learning algorithms to deeply mine the data and build detailed user portraits; The AI ​​recommendation engine uses deep learning technology to provide personalized medical services or product recommendations based on user portraits and historical interaction data. The recommended content covers disease prevention, health management, and disease treatment. The customer interaction platform supports multiple communication methods, including instant messaging, telephone and email, to promote effective communication between medical staff and patients; The effect evaluation and feedback mechanism quantifies the effect of marketing activities by setting a KPI indicator system, and continuously optimizes the recommendation model and service process based on user feedback.

2. The SCRM-based medical intelligent marketing system according to claim 1 is characterized by: The data collection module uses the official API interface provided by major social media platforms to guide users through the OAuth2.0 authorization protocol, authorize applications to access their public information, and then use the HTTP request library to call the API interface to obtain the user's public information; at the same time, it is compatible with third-party health application interfaces, and is compatible with the relevant data information transmitted and obtained by the interfaces of third-party applications such as wearable devices and health management apps, so as to obtain more comprehensive patient health data through the same authorized access method; when verifying data, a hash function is used to calculate the data summary, and the summary of the two data sources is compared to see if they are consistent. The sensitive data is encrypted using symmetric encryption technology. The encryption formula is divided into four steps. Step 1: initial round key addition, the implementation formula is: In the formula, P is the plaintext block (128 bits), K0 is the initial key (128 bits), and C0 is the output of the initial round; Step 2: Multiple rounds of transformation, the implementation formula is: S i,j =S box [S i,j ] In the formula, S i,j is the byte in row i and column j of the state matrix, S box is the S-box table; Step 3: Final round transformation, the implementation formula is: In the formula, W[i] is the i-th word, N k The word number T(W[i-1]) of the key is a transformation function, including byte replacement, circular shift and round constant XOR. The symmetric encryption algorithm is used to achieve efficient encryption and decryption of data. These transformations include byte replacement, row shift, column mixing and key addition steps, ensuring the high security and anti-attack capability of the algorithm.

3. The SCRM-based medical intelligent marketing system according to claim 1 is characterized by: The data processing and analysis module uses natural language processing and machine learning algorithms to parse unstructured data, uses the TF-IDF method to extract text features and then selects the support vector machine SVM machine learning algorithm for training, and the implementation formula is: TF-IDF(t,d)=TF(t,d)×IDF(t) In the formula, TF(t,d) is the frequency of word t appearing in document d, and IDF(t) is the inverse document frequency, calculated as: In the formula, N is the total number of documents, n t is the number of documents containing word t; the trained model is used to identify the potential needs of patients and possible health risks, which are valuable features for marketing decisions; at the same time, sentiment analysis technology is introduced to select the naive Bayes model for model training through machine learning methods, and the implementation formula is: In the formula, P(C k |x) is the category C under the given feature vector x k The posterior probability, P(x|C k ) is Category C k The conditional probability of the eigenvector x under k ) is Category C k The prior probability of P(x) is the marginal probability of the feature vector x. The trained model is used to classify the sentiment of the new text, judge the positive, negative and neutral sentiment, evaluate the patient's attitude towards specific services and products, and assist in the formulation of marketing strategies. The graph neural network technology is used to build a relationship network between users, and the GCN model is used to learn the embedded representation of the node. The implementation formula is: In the formula, H (l+1) is the node feature matrix of the lth layer, is the adjacency matrix with self-loops added, yes The degree matrix, W (l) is the weight matrix of the lth layer, σ is the activation function, which discovers the potential community effect and provides support for group marketing; these technical means and implementation steps together constitute an efficient and intelligent data processing and analysis system, providing strong support for other modules of the medical intelligent marketing system.

4. The SCRM-based medical intelligent marketing system according to claim 1 is characterized by: The AI ​​recommendation engine module can dynamically adjust the recommended content according to the real-time needs of users. First, the user's search history and page browsing history are collected through the front-end application and the website. When it is detected that the user searches for relevant information about a certain disease, the recommendation system is triggered to generate new recommended content. Then, according to the user's search keywords, relevant popular science articles and expert lecture videos are matched from the content library, and the matched content is pushed to the user through pop-up windows and notification bars. The formula for calculating the similarity matching between keywords and content libraries is as follows: In the formula, A and B are two vectors, representing the TF-IDF vectors of two documents respectively, · represents the vector dot product, ∥A∥ and ∥B∥ respectively represent the modulus of vectors A and B; obtain the user's geographic location information through the user's IP address and GPS positioning technology, update the user portrait based on the user's geographic location and time factors, and add features of geographic and time dimensions. The collaborative filtering formula is used to generate recommendations based on the user's historical behavior and the behavior of similar users. The implementation formula is: In the formula, is the predicted rating of user u for item i, is the average rating of user u, sin(u,v) is the similarity between user u and user v, r vi is the actual rating of item i by user v, is the average rating of user v, N(u,i) is the set of users similar to user u who have rated item i, and the generated recommended content is pushed to the user; the recommendation engine detects the user's browser settings and the language preference set by the user through NLP technology, and uses the machine translation model Transformer to translate the recommended content into the language selected by the user. The implementation formula is: In the formula, Q is the query vector, K is the key vector, V is the value vector, and d k is the dimension of the key vector, softmax is the softmax function, which is used to normalize the probability distribution. Multi-language switching is supported in the front-end application, and users can choose different language versions to meet the needs of international patients.

5. The SCRM-based medical intelligent marketing system according to claim 1 is characterized by: The customer interaction platform module has intelligent customer service functions, collects and organizes common questions and their answers, builds a knowledge base, uses support vector machines (SVMs) for intent recognition, and identifies user intent. The implementation formula is: In the formula, y is the user's intention, x is the user's input text, Y is the set of all possible intentions, and P(y|x) is the probability of intention y given the input text x. At the same time, the module supports automatic speech recognition to realize the recognition and understanding of speech input. The front end converts the user's speech input into text, cleans and pre-processes the converted text, uses the speech recognition formula to identify the user's intention and extract key entities from the text, and finally selects the appropriate answer and performs the corresponding operation based on the user's intention and entity. The implementation formula is: In the formula, is the recognized text, X is the input speech signal, P(w|X) is the probability of text w given the speech signal X; the implementation formula for entity extraction is: In the formula, y i is the entity tag of the i-th word, x i is the i-th word, x i-1 and x i+1 are adjacent words, Y is the set of all possible entity tags, P(y i |x i ,x i-1 ,x i+1 ) is the probability of an entity label given a word and its context; through intelligent customer service supporting speech recognition and natural language understanding technology, it can accurately capture the user's intention and make appropriate responses.

6. The SCRM-based medical intelligent marketing system according to claim 1 is characterized by: The effect evaluation and feedback mechanism module defines a series of KPI indicators according to the marketing objectives, including click-through rate, conversion rate, and follow-up rate. First, the user's click count, purchase count, and follow-up count data are collected through the front-end application, and the collected data is cleaned, deduplicated, and format converted to ensure data quality. Then, the values ​​of various indicators are calculated according to the defined KPI formula. Finally, the calculation results are displayed in the form of charts and dashboards for easy viewing and analysis by the management and marketing team. At the same time, the effects of different marketing strategies are compared through A / B testing to find the best solution. The module randomly divides users into multiple groups, each group uses different marketing strategies, and runs marketing activities of multiple groups at the same time in the same environment, records the user behavior data of each group, uses statistical methods to analyze the data of each group, evaluates the effects of different strategies, and selects the best performing strategy as the final solution based on the data analysis results. The chi-square test is used to compare whether there are significant differences in the categorical variables of multiple groups. The implementation formula is: In the formula, O i is the observed value, E i is the expected value, n is the number of categorical variables; the last module will regularly collect and summarize the data of marketing activities, conduct in-depth analysis of the summarized data, identify the reasons for success and failure, and write detailed marketing reports, including data charts, analysis conclusions, and recommended measures. The marketing report will be issued to relevant management and marketing teams through emails and meetings, and feedback from relevant personnel will be collected. The report content and analysis methods will be continuously improved, lessons learned will be summarized, and future marketing work will be guided.

7. The SCRM-based medical intelligent marketing system according to claim 1 is characterized in that: The system comprises the following steps: S1. Obtain patient-related information through the data collection module to ensure that the data source is legal and compliant; S2. Generate user portraits using data processing and analysis modules, including basic information, health status, and consumption habits; S3. AI recommendation engines provide personalized services or product recommendations based on user portraits. Recommended content must undergo strict review to ensure its scientificity and legality. S4. Realize doctor-patient communication through the customer interaction platform and provide convenient consulting services; S5. Use the effect evaluation and feedback mechanism to monitor the effectiveness of marketing activities, adjust strategies in a timely manner, and improve marketing efficiency.

8. The SCRM-based medical intelligent marketing system according to claim 7 is characterized by: In step S1: the data collection process strictly abides by the relevant laws and regulations of the "Cybersecurity Law" and the "Personal Information Protection Law" to ensure information security and personal privacy protection; and clearly informs users of the purpose, scope and usage rules of data collection, and respects users' right to know and right to choose.

9. The SCRM-based medical intelligent marketing system according to claim 7 is characterized by: In step S3: personalized recommendations not only take into account user preferences, but also take into account the professionalism and safety of medical services to avoid damaging the interests of patients due to excessive marketing; for special groups (such as minors and the elderly), recommendations will be more cautious to prevent misleading or adverse effects; and medical experts are regularly invited to review the recommendations to ensure that they comply with the latest medical standards and ethical norms.

10. The SCRM-based medical intelligent marketing system according to claim 7, characterized in that: In step S5: In addition to the routine marketing effectiveness evaluation, it also includes a survey on patient satisfaction as an important reference for optimizing services; conducting regular user surveys to collect feedback on system functions, interface design, ease of operation, etc., and continuously improving user experience; establishing a sound complaint handling mechanism, actively responding to users' reasonable demands, and maintaining a good brand image.

Images