Behavior analysis and medical service pushing method and system based on big data

Abstract

The invention belongs to the technical field of big data, and particularly relates to a behavior analysis and medical service pushing method and system based on big data, and the method comprises the steps: carrying out the matching of each piece of medical service information with a medical service dictionary, and obtaining the feature vocabularies of each piece of medical service information; based on the simultaneous occurrence frequency of different feature vocabularies of the medical service information, establishing an association network of a medical service dictionary; obtaining the credibility of each piece of medical service information based on the range and aggregation condition of the feature vocabulary of each piece of medical service information in the association network of the medical service dictionary; obtaining a positive emotion index and a negative emotion index of each piece of medical service information based on the emotion vocabulary expression of each piece of medical service information; and obtaining the credibility of each medical service noun, and obtaining to-be-pushed medical service information of each user by using a collaborative filtering algorithm. According to the invention, the information quality and the pushing effectiveness of medical service pushing by using the collaborative filtering algorithm are improved.

Description

Technical Field

[0001] The present invention relates to the field of big data technology. More specifically, the present invention relates to a method and system for behavior analysis and medical service push based on big data. Background Art

[0002] With the development of technology and the increasing attention people pay to health, medical service platforms have gradually developed. People can retrieve relevant medical information through medical service platforms, thereby increasing their understanding of medical services related to their own diseases or diseases of interest. However, only obtaining the information required by users through user retrieval has relatively high requirements for the user's retrieval ability and requires users to judge the quality of the information obtained. Therefore, combining big data to analyze users' behaviors and perform personalized medical service push can improve the efficiency of users obtaining relevant medical services.

[0003] In order to accurately push relevant information to users, the collaborative filtering algorithm is often used at present to obtain the received information of similar behavior users of the target user and push it to the target user. In related technologies, for example, the Chinese patent document with the authorization announcement number CN109582875B discloses a method and system for personalized recommendation of online medical education resources, which discloses establishing a case recommendation model using a content-based collaborative filtering recommendation algorithm to generate recommendation results for the target user, reducing the time consumed by the target user's active search, and reducing the user's learning cost. The Chinese patent document with the authorization announcement number CN118277668B discloses a method and system for information push for financial services, which discloses combining the collaborative filtering algorithm to obtain the predicted scores of each user for each financial service, and obtaining the financial service recommendation list corresponding to each user to achieve financial service information push.

[0004] In the process of using the collaborative filtering algorithm for medical service push, there is a problem of unstable quality of medical service information. Information that can arouse people's emotions often has a high popularity, but lacks necessity and professionalism. Therefore, the collaborative filtering algorithm may recommend medical service information that does not conform to the user's condition to a large number of people who cannot distinguish medical information, affecting users' access to effective medical service information and even affecting disease treatment. Summary of the Invention

[0005] To solve the technical problem of unstable quality of medical service information when using the collaborative filtering algorithm for medical service information push, the present invention provides solutions in the following aspects.

[0006] In a first aspect, the present invention provides a method for behavior analysis and medical service push based on big data, including: Build a medical service dictionary, where the medical service dictionary contains a number of disease names, a number of symptom names, and a number of medical service names; obtain a number of medical service information; obtain the medical service demand sets of a number of users, and the medical service demand sets of users are all subsets of the medical service dictionary; match each medical service information with the medical service dictionary to obtain the characteristic vocabulary of each medical service information; based on the frequency of simultaneous occurrence of different characteristic vocabulary of medical service information, establish an association network of the medical service dictionary; based on the scope and aggregation of the characteristic vocabulary of each medical service information in the association network of the medical service dictionary, obtain the credibility of each medical service information; based on the emotional vocabulary performance of each medical service information, obtain the positive emotional index and negative emotional index of each medical service information; according to the positive emotional index and negative emotional index of the medical service information related to each medical service name, and the credibility of the medical service information related to each medical service name, obtain the credibility of each medical service name; based on the medical service demand sets of users and the credibility of each medical service name, use the collaborative filtering algorithm to obtain the medical service information to be pushed for each user.

[0007] The present invention screens the medical service information pushed to users based on credibility, improving the quality and efficiency of the medical service information that users can obtain. The present invention obtains the effectiveness of medical service information and medical service vocabulary based on emotional vocabulary performance and medical service-related vocabulary performance, making the quality judgment of medical service information more accurate, thereby avoiding a large number of low-quality information being pushed to users through the collaborative filtering algorithm.

[0008] Preferably, the establishment of the association network of the medical service dictionary based on the frequency of simultaneous occurrence of different characteristic vocabulary of medical service information includes: performing a union operation on the characteristic vocabulary of all medical service information to obtain the characteristic vocabulary set and the type of characteristic vocabulary of medical service information; obtaining the relevance of any two characteristic vocabulary according to the frequency of simultaneous occurrence of different characteristic vocabulary of medical service information; taking the inverse normalization result of the relevance of any two characteristic vocabulary as the distance between the two characteristic vocabulary in the association network of the medical service dictionary, taking all characteristic vocabulary as the nodes of the network, and connecting all characteristic vocabulary with a distance greater than 0 to obtain the association network of the medical service dictionary.

[0009] The present invention establishes an association network of the medical service dictionary, enabling the association of medical service-related vocabulary to be visually represented, thereby enabling the determination of the association of each vocabulary in the medical service dictionary.

[0010] Preferably, the relevance of any two characteristic vocabulary satisfies the expression: ; In the formula, Indicates the relevance between the i-th characteristic word and the c-th characteristic word; , Indicates the number of medical service information containing the i-th characteristic word and the c-th characteristic word; Indicates the number of medical service information containing both the i-th characteristic word and the c-th characteristic word; Indicates the absolute value function; Indicates the normalization function; Indicates the exponential function with the natural constant as the base.

[0011] The present invention obtains the relevance of different characteristic words through the co-occurrence of different characteristic words, ensuring the accuracy of the analysis of the association of characteristic words.

[0012] Preferably, the obtaining of the credibility of each medical service information includes: Obtaining the number of words of each medical service information, obtaining the number of words of each characteristic word of each medical service information, and obtaining the types of characteristic words included in each medical service information; ; In the formula, Indicates the credibility of the a-th medical service information; Indicates the set of the number of words of the characteristic words of the a-th medical service information; Indicates the number of words of the a-th medical service information; Indicates the types of characteristic words included in the a-th medical service information; Indicates the distance between the b-th characteristic word and the d-th characteristic word included in the a-th medical service information in the association network of the medical service dictionary; Indicates the exponential function with the natural constant as the base.

[0013] Preferably, the obtaining of the positive sentiment index and the negative sentiment index of each medical service information includes: obtaining the set of sentiment words of each medical service information according to the performance of the sentiment words of each medical service information; obtaining the relative sentiment relationship of different sentiment words based on the voting mechanism; obtaining the final sentiment degree of each sentiment word according to the relative sentiment relationship of different sentiment words; adding up the final sentiment degrees of all sentiment words with a final sentiment degree greater than 0 of the a-th medical service information, which is recorded as the positive sentiment index of the a-th medical service information; adding up the final sentiment degrees of all sentiment words with a final sentiment degree less than 0 of the a-th medical service information and taking the absolute value, which is recorded as the negative sentiment index of the a-th medical service information.

[0014] The present invention calculates the positive sentiment index and the negative sentiment index, providing a clear basis for analyzing the sentiment performance of medical service information, so as to more accurately obtain the credibility of medical service words.

[0015] Preferably, the emotional vocabulary set of each medical service information is obtained based on the emotional vocabulary expression of each medical service information, including: using jieba participle to perform part-of-speech tagging on each medical service information, obtaining words with the part of speech of adjective or adverb in each medical service information, and recording them as the emotional vocabulary of each medical service information; performing a union operation on the emotional vocabulary of each medical service information to obtain the emotional vocabulary set of each medical service information.

[0016] Preferably, the voting mechanism is based on obtaining the relative emotional relationship of different emotional words, including: performing a union operation on the emotional words of each medical service information to obtain an emotional word set of each medical service information; setting up a voting mechanism involving several people, for any two emotional words, each person votes for the emotional word that he or she personally thinks is more positive among the two emotional words, and the emotional word with the highest number of votes among the two emotional words is recorded as the relatively positive emotional word among the two emotional words, and the emotional word with the lowest number of votes among the two emotional words is recorded as the relatively negative emotional word among the two emotional words.

[0017] The present invention determines the emotion levels of different emotion words through a voting mechanism, improves the robustness of using emotion word sequences, and improves the accuracy of emotion analysis of medical service information.

[0018] Preferably, the method of obtaining the final emotion degree of each emotion word according to the emotion relative relationship of different emotion words includes: constructing an emotion word degree tree according to the emotion relative relationship of different emotion words, wherein the emotion word degree tree is a binary sorted tree; performing an in-order traversal on the emotion word degree tree to obtain an in-order emotion word sequence; recording the emotion word corresponding to the ordinal median of the in-order emotion word sequence as a neutral emotion word, subtracting the ordinal number of the neutral emotion word from the ordinal number of the u-th emotion word in the in-order emotion word sequence, and then dividing the result by the maximum ordinal number of the in-order emotion word sequence to record the final emotion degree of the u-th emotion word.

[0019] Preferably, obtaining the credibility of each medical service noun includes: Obtain all medical service information containing the s-th medical service noun, recorded as the relevant medical service information of the s-th medical service noun; ; In the formula, represents the credibility of the sth medical service noun; represents the number of related medical service information of the s-th medical service noun; represents the credibility of the ath related medical service information of the sth medical service noun; , Denote the positive sentiment index and negative sentiment index of the ath relevant medical service information of the sth medical service term; Denote the absolute value function; Denote the exponential function with the natural constant as the base.

[0020] In a second aspect, the present invention provides a behavior analysis and medical service push system based on big data, including a processor and a memory. The memory stores computer program instructions, and when the computer program instructions are executed by the processor, the above-mentioned behavior analysis and medical service push method based on big data is implemented.

[0021] By adopting the above technical solution, the above-mentioned behavior analysis and medical service push method based on big data is generated into a computer program and stored in the memory to be loaded and executed by the processor, so as to manufacture a terminal device according to the memory and the processor, which is convenient to use.

[0022] The beneficial effects of the present invention are as follows: (1) Based on big data technology, the present invention obtains the credibility of medical service terms through medical service information related to medical service vocabulary, making the acquisition of the quality of medical service information more accurate, thus providing a basis for accurately pushing medical services for users using the collaborative filtering algorithm; (2) The final sentiment degree of the sentiment vocabulary of the present invention provides a basis for analyzing medical service information from the sentiment aspect, avoiding the problem that the collaborative filtering algorithm may push low-quality medical service information to users. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 is a flowchart schematically showing a behavior analysis and medical service push method based on big data in the present invention; Figure 2 is a schematic diagram schematically showing the association network of the medical service dictionary. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0024] An embodiment of the present invention discloses a behavior analysis and medical service push method based on big data, referring to Figure 1 , including steps S1 - S4: S1: Establish a medical service dictionary, where the medical service dictionary includes several disease terms, several symptom terms, and several medical service terms; obtain several medical service information; obtain the medical service demand set of several users.

[0025] It should be noted that taking a social media platform as an example, the platform will use a collaborative filtering algorithm based on users to find similar user groups and recommend relevant content according to the preferences of similar users, which can improve the utilization rate of the platform by users and reduce the retrieval cost of users. In order to increase the access volume of the pushed content, users who send the pushed content often add words that can trigger people's emotional changes. However, when it comes to the recommendation of medical service information, more attention needs to be paid to the information quality of medical service information. When low-quality medical service information is spread more widely by the collaborative filtering algorithm, it is easy to lead to the spread of incorrect information. In order to ensure the information quality of medical service push, the present invention first obtains relevant push information of various medical services, then analyzes the credibility of each medical service information, and finally pushes the medical service information with high credibility and relevant to the user to be pushed to the user to be pushed.

[0026] Specifically, a medical service dictionary is established manually, and the medical service dictionary includes a number of disease nouns, a number of symptom nouns, and a number of medical service nouns.

[0027] On the content platform involving medical service push, a number of medical service information is obtained.

[0028] Obtain the search records of a number of users, use the maximum matching word segmentation algorithm based on the dictionary to segment the search records of all users, and the dictionary used is the medical service dictionary, to obtain a number of medical service demand vocabulary of each user, and form the medical service demand set of each user. It should be noted that the search records of users may include diseases, symptoms, and medical services, so the medical service demand set of users may include disease nouns, symptom nouns, and medical service nouns.

[0029] So far, the medical service dictionary, a number of medical service information, and the medical service demand sets of a number of users have been obtained.

[0030] S2: Match each medical service information with the medical service dictionary to obtain the characteristic vocabulary of each medical service information; establish an association network of the medical service dictionary based on the frequency of simultaneous appearance of different characteristic vocabulary of the medical service information; obtain the credibility of each medical service information based on the scope and aggregation of the characteristic vocabulary of each medical service information in the association network of the medical service dictionary.

[0031] It should be noted that among all medical service information, due to the differences in the medical capabilities of senders, the medical service information also shows quality differences, and in order to be easily spread, some medical service information contains non-professional terms and non-authentic content. Therefore, it is necessary to judge the credibility of the obtained medical service information.

[0032] It should be further noted that a disease often has multiple symptoms and multiple medical services. Taking heart disease as an example, medical services include monitoring heartbeats through wearable devices, non-invasive screening, bypass surgery for multi-vessel lesions, etc. Different medical services are for different symptoms and purposes. Therefore, symptoms and medical services often have a correlation. When the medical service information contains more descriptions of medical services and there are associations in the descriptions of these medical service information, it indicates that the medical service information reflects the correlation between symptoms and medical services and has a higher credibility. Therefore, the present invention analyzes the credibility of each medical service information in combination with the correlation of the medical service-related vocabulary contained in the medical service information.

[0033] Specifically, each medical service information is matched with a medical service dictionary to obtain the characteristic vocabulary of each medical service information: Use the maximum matching word segmentation algorithm based on the dictionary to segment each medical service information, and record the obtained vocabulary as the characteristic vocabulary of the target medical service information. The dictionary used is the medical service dictionary. It should be noted that since the medical service dictionary is used in the word segmentation process, all characteristic vocabulary is included in the medical service dictionary.

[0034] It should be noted that there is a correlation between diseases, symptoms and medical services. For example, antiplatelet drugs are used to prevent the formation of blood clots in heart disease. The characteristic vocabulary contained in a medical service information reflects the correlation between diseases, symptoms and medical services, and the higher the frequency of simultaneous occurrence of two characteristic vocabulary in all medical service information, the higher the correlation between the two characteristic vocabulary. Based on the correlation between characteristic vocabulary, an irregular medical service dictionary can be turned into a network reflecting the relationship between each vocabulary in the medical service dictionary.

[0035] Preferably, an association network of the medical service dictionary is established based on the frequency of simultaneous occurrence of different characteristic vocabulary in the medical service information: Perform a union operation on the characteristic vocabulary of all medical service information to obtain the characteristic vocabulary set and the type of characteristic vocabulary of the medical service information.

[0036] It should be noted that the more frequently the i-th type of characteristic vocabulary and the c-th type of characteristic vocabulary appear simultaneously, and the closer the number of medical service information containing the i-th type of characteristic vocabulary and the c-th type of characteristic vocabulary, the stronger the correlation between the i-th type of characteristic vocabulary and the c-th type of characteristic vocabulary.

[0037] The correlation between any two types of characteristic vocabulary satisfies the expression: ; In the formula, represents the correlation between the i-th type of characteristic vocabulary and the c-th type of characteristic vocabulary; , represents the quantity of medical service information containing the \(i\)-th characteristic term and the \(c\)-th characteristic term; represents the quantity of medical service information containing both the \(i\)-th characteristic term and the \(c\)-th characteristic term; represents the absolute value function; represents the normalization function; represents the exponential function with the natural constant as the base.

[0038] In the formula, represents the proportion of medical service information containing the \(c\)-th characteristic term among the medical service information containing the \(i\)-th characteristic term; represents the proportion of medical service information containing the \(i\)-th characteristic term among the medical service information containing the \(c\)-th characteristic term; represents the product of the proportions of the \(i\)-th characteristic term and the \(c\)-th characteristic term respectively in the medical service information containing the other characteristic term. The larger this value is, the greater the proportion of the simultaneous occurrence of the \(i\)-th characteristic term and the \(c\)-th characteristic term; represents the difference in the quantity of medical service information containing the \(i\)-th characteristic term and the \(c\)-th characteristic term. The smaller this value is, the closer the quantities of medical service information containing the \(i\)-th characteristic term and the \(c\)-th characteristic term are. At the same time, if the proportion of the simultaneous occurrence of the \(i\)-th characteristic term and the \(c\)-th characteristic term is larger, it indicates that the probability of the mere simultaneous occurrence of the \(i\)-th characteristic term and the \(c\)-th characteristic term is greater. Therefore, the stronger the correlation between the \(i\)-th characteristic term and the \(c\)-th characteristic term.

[0039] Take the inverse proportional normalization result of the correlation between any two characteristic terms as the distance between the two characteristic terms in the association network of the medical service dictionary. Take all characteristic terms as the nodes of the network, and connect all characteristic terms with a distance greater than 0 to obtain the association network of the medical service dictionary.

[0040] Thus, the association network of the medical service dictionary is obtained.

[0041] It should be noted that as Figure 2 is a schematic diagram of the association network of the medical service dictionary. Due to different fields, the association network of the medical service dictionary often shows multiple aggregation regions. One aggregation region represents the medical service-related vocabulary of a disease or a field. Therefore, the more concentrated the characteristic terms contained in the medical service information are in the association network of the medical service dictionary, and the higher the proportion of the number of characters of the characteristic terms in the medical service information, the stronger the professionalism of the medical service information, and thus the higher the credibility.

[0042] Preferably, based on the proportion of the number of characters of the characteristic terms in each medical service information and the aggregation situation in the association network of the medical service dictionary, obtain the credibility of each medical service information: Obtain the word count of each medical service information, obtain the word count of each characteristic word of each medical service information, and obtain the types of characteristic words included in each medical service information.

[0043] The credibility of any medical service information satisfies the expression: ; In the formula, represents the credibility of the a-th medical service information; represents the set of word counts of the characteristic words of the a-th medical service information; represents the word count of the a-th medical service information; represents the types of characteristic words included in the a-th medical service information; represents the distance between the b-th and d-th characteristic words included in the a-th medical service information in the association network of the medical service dictionary; represents the exponential function with the natural constant as the base.

[0044] In the formula, represents the proportion of the word count of the characteristic words in the a-th medical service information. The larger this value is, the more content related to the medical service the medical service information has, the stronger the relevance, and thus the higher the credibility; represents the average distance of all the characteristic words included in the a-th medical service information in the association network of the medical service dictionary. The larger this value is, the lower the relevance of all the characteristic words included in the a-th medical service information, indicating that the a-th medical service information does not revolve around a medical entity, and thus the credibility of the a-th medical service information is lower.

[0045] Thus, the credibility of each medical service information has been obtained.

[0046] S3: Based on the emotional word performance of each medical service information, obtain the positive emotional index and negative emotional index of each medical service information; according to the positive emotional index and negative emotional index of the medical service information related to each medical service noun, and the credibility of the medical service information related to each medical service noun, obtain the credibility of each medical service noun.

[0047] It should be noted that when transmitting correct medical service information, for example, in papers, emotional words such as "touched" are often avoided, while in marketing and promotion, for example, advertisements will use emotional words to increase the possibility of successful promotion of medical services. Positive emotional words promote the promotion of medical services, and negative emotional words are used to prevent users from using other medical services. Therefore, the stronger the emotional word performance in the medical service information and the greater the positive-negative emotional difference, the lower the credibility of the medical service information.

[0048] It should be further noted that emotional words often appear as adjectives and adverbs. Therefore, by obtaining adjectives and adverbs in medical service information and performing emotional annotation, the positive and negative emotional analysis of medical service information can be completed. Natural language processing technology can efficiently process text and perform part-of-speech tagging on the text. Therefore, the present invention uses natural language processing technology to extract adjectives and adverbs, and manually sorts the extracted adjectives and adverbs according to their emotional degrees, so as to obtain the positive emotional index and negative emotional index of each medical service information based on the emotional degrees of all adjectives and adverbs in each medical service information.

[0049] Specifically, based on the emotional word performance of each medical service information, the positive emotional index and negative emotional index of each medical service information are obtained: Use jieba word segmentation to perform part-of-speech tagging on each medical service information, and obtain the words with the part-of-speech of adjectives or adverbs in each medical service information, which are recorded as the emotional words of each medical service information. It should be noted that jieba word segmentation is an existing Chinese word segmentation tool that can perform word segmentation and part-of-speech tagging on text.

[0050] Perform union operation on the emotional words of each medical service information to obtain the emotional word set of each medical service information.

[0051] Set up a voting mechanism including several people. For any two emotional words, each person votes for the emotional word that they think is more positive among the two emotional words. The emotional word with the highest number of votes among the two emotional words is recorded as the relatively positive emotional word among the two emotional words, and the emotional word with the lowest number of votes among the two emotional words is recorded as the relatively negative emotional word among the two emotional words. It should be noted that since the emotional degrees of emotional words may vary for different people, the voting mechanism can ensure more accurate analysis of the emotional degrees of emotional words.

[0052] Construct an emotional degree function, and the emotional degree function satisfies the expression: ; In the formula, represents the emotional degree function of the u-th emotional word and the v-th emotional word, and f(u) and f(v) represent the relative emotional degrees of the u-th emotional word and the v-th emotional word among the two emotional words; when the u-th emotional word is the relatively positive emotional word among the u-th emotional word and the v-th emotional word, f(u) is 1 and f(v) is -1; when the u-th emotional word is the relatively negative emotional word among the two emotional words, f(u) is -1 and f(v) is 1.

[0053] For the set of sentiment words, take the first sentiment word as the root node. If the sentiment degree function of the second sentiment word and the first sentiment word is , then place the second sentiment word in the right subtree of the node corresponding to the first sentiment word; if the sentiment degree function of the second sentiment word and the first sentiment word is , then place the second sentiment word in the left subtree of the node corresponding to the first sentiment word; place all sentiment words in turn to form a binary sort tree, denoted as the sentiment word degree tree. It should be noted that in the sentiment word degree, the relative sentiment degrees of all nodes in the left subtree are less than the value of the root node, and the relative sentiment degrees of all nodes in the right subtree are greater than the value of the root node. And performing an in-order traversal of the sentiment word degree tree, the obtained in-order sentiment word sequence is a relationship sequence with increasing sentiment degree.

[0054] Perform an in-order traversal of the sentiment word degree tree to obtain an in-order sentiment word sequence.

[0055] Denote the sentiment word corresponding to the ordinal median of the in-order sentiment word sequence as the neutral sentiment word. Subtract the ordinal number of the neutral sentiment word from the ordinal number of the u-th sentiment word in the in-order sentiment word sequence, and then divide by the maximum ordinal number of the in-order sentiment word sequence, which is denoted as the final sentiment degree of the u-th sentiment word.

[0056] Add up the final sentiment degrees of all sentiment words with a final sentiment degree greater than 0 in the a-th medical service information, which is denoted as the positive sentiment index of the a-th medical service information; add up the final sentiment degrees of all sentiment words with a final sentiment degree less than 0 in the a-th medical service information, and take the absolute value, which is denoted as the negative sentiment index of the a-th medical service information.

[0057] So far, the positive sentiment index and negative sentiment index of each medical service information have been obtained.

[0058] It should be noted that the larger the positive sentiment index and negative sentiment index of the medical service information related to the medical service noun, the lower the credibility of the medical service information related to the medical service noun, and thus the lower the credibility of the medical service noun; in addition, considering that medical service senders will process medical service pushes with a relatively high negative sentiment index due to interest relationships to avoid the spread of bad information, therefore, in the medical service information related to the medical service noun, the greater the overall difference between the positive sentiment index and the negative sentiment index, the lower the credibility of the medical service noun; at the same time, the lower the overall credibility of the medical service information related to the medical service noun, the lower the credibility of the medical service noun.

[0059] Preferably, according to the positive sentiment index and negative sentiment index of medical service information related to each medical service term, and the credibility of medical service information related to each medical service term, obtain the credibility of each medical service term: Obtain all medical service information containing the s-th medical service term, denoted as the medical service information related to the s-th medical service term, and obtain the positive sentiment index and negative sentiment index of all medical service information related to the s-th medical service term.

[0060] The credibility of any medical service term satisfies the expression: ; In the formula, represents the credibility of the s-th medical service term; represents the quantity of medical service information related to the s-th medical service term; represents the credibility of the a-th medical service information related to the s-th medical service term; and represent the positive sentiment index and negative sentiment index of the a-th medical service information related to the s-th medical service term; represents the absolute value function; represents the exponential function with the natural constant as the base.

[0061] In the formula, represents the influence of the a-th medical service information related to the s-th medical service term on the credibility of the s-th medical service term. The greater the credibility of the a-th medical service information related to the s-th medical service term, the smaller the positive sentiment index and negative sentiment index, and the smaller the difference between the positive sentiment index and negative sentiment index, the greater the influence on the credibility of the s-th medical service term, making the credibility of the s-th medical service term greater.

[0062] Thus, the credibility of each medical service term is obtained.

[0063] S4: Based on the set of users' medical service requirements and the credibility of each medical service term, use the collaborative filtering algorithm to obtain the medical service information to be pushed for each user.

[0064] Specifically, any user is denoted as the target user. On the association network of the medical service dictionary, all the medical service terms that can be reached by the terms in the medical service demand set of the target user are obtained, which are denoted as the relevant medical service terms of the target user. Among the relevant medical service information of the relevant medical service terms of the target user, in the order of descending credibility, the Top-N algorithm is used to obtain the medical service information to be pushed for all the relevant medical service terms of the target user, and the target user is pushed in the descending order of all the medical service information to be pushed. It should be noted that the number of medical service information to be pushed selected by the Top-N algorithm, that is, the N value, is set by the implementer according to the actual implementation situation. For example, the N value can be set to 3.

[0065] Thus, the medical service push for each user is completed.

[0066] The embodiment of the present invention also discloses a big data-based behavior analysis and medical service push system, including a processor and a memory. The memory stores computer program instructions, and when the computer program instructions are executed by the processor, a big data-based behavior analysis and medical service push method according to the present invention is implemented.

[0067] The above system also includes other components well-known to those skilled in the art such as a communication bus and a communication interface. Their settings and functions are known in the art, so they will not be described in detail here.

[0068] Although this specification has shown and described multiple embodiments of the present invention, it is obvious to those skilled in the art that such embodiments are provided only by way of example. Those skilled in the art will think of many changes, alterations, and alternative ways without departing from the spirit and idea of the present invention. It should be understood that various alternative solutions to the embodiments of the present invention described herein can be adopted in the process of practicing the present invention.

Claims

1. A method for behavior analysis and medical service push based on big data, characterized in that, Including: Establish a medical service dictionary, where the medical service dictionary contains a number of disease names, a number of symptom names, and a number of medical service names; obtain a number of medical service information; obtain the medical service demand sets of a number of users, and the medical service demand sets of users are all subsets of the medical service dictionary; Match each medical service information with the medical service dictionary to obtain the characteristic vocabulary of each medical service information; establish an association network of the medical service dictionary based on the frequency of simultaneous occurrence of different characteristic vocabulary of the medical service information; obtain the credibility of each medical service information based on the scope and aggregation of the characteristic vocabulary of each medical service information in the association network of the medical service dictionary; Obtain the positive sentiment index and negative sentiment index of each medical service information based on the sentiment vocabulary performance of each medical service information; obtain the credibility of each medical service name according to the positive sentiment index and negative sentiment index of the medical service information related to each medical service name, and the credibility of the medical service information related to each medical service name; Based on the medical service demand sets of users and the credibility of each medical service name, use the collaborative filtering algorithm to obtain the medical service information to be pushed for each user.

2. The method for behavior analysis and medical service push based on big data according to claim 1, wherein The establishing of the association network of the medical service dictionary based on the frequency of simultaneous occurrence of different characteristic vocabulary of the medical service information includes: Perform a union operation on the characteristic vocabulary of all medical service information to obtain the characteristic vocabulary set and the type of characteristic vocabulary of the medical service information; obtain the relevance between any two characteristic vocabulary according to the frequency of simultaneous occurrence of different characteristic vocabulary of the medical service information; Take the inverse normalization result of the relevance between any two characteristic vocabulary as the distance between the two characteristic vocabulary in the association network of the medical service dictionary, take all characteristic vocabulary as the nodes of the network, and connect all characteristic vocabulary with a distance greater than 0 to obtain the association network of the medical service dictionary.

3. The method for behavior analysis and medical service push based on big data according to claim 2, characterized in that The relevance between any two characteristic vocabulary satisfies the expression: ； In the formula, represents the relevance between the i-th characteristic word and the c-th characteristic word; , represents the quantity of medical service information containing the i-th characteristic word and the c-th characteristic word; represents the quantity of medical service information containing both the i-th characteristic word and the c-th characteristic word; represents the absolute value function; represents the normalization function; represents the exponential function with the natural constant as the base.

4. A method for behavior analysis and medical service push based on big data according to claim 1, characterized in that The obtaining of the credibility of each medical service information includes: Obtain the number of words of each medical service information, obtain the number of words of each characteristic vocabulary of each medical service information, and obtain the type of characteristic vocabulary included in each medical service information; ； In the formula, represents the credibility of the a-th medical service information; represents the set of the number of words of the characteristic vocabulary of the a-th medical service information; represents the number of words of the a-th medical service information; represents the type of characteristic vocabulary included in the a-th medical service information; represents the distance between the b-th characteristic vocabulary and the d-th characteristic vocabulary included in the a-th medical service information in the association network of the medical service dictionary; represents the exponential function with the natural constant as the base.

5. A method for behavior analysis and medical service push based on big data according to claim 1, characterized in that, The obtaining of the positive sentiment index and negative sentiment index of each medical service information includes: According to the sentiment vocabulary performance of each medical service information, obtain the sentiment vocabulary set of each medical service information; based on the voting mechanism, obtain the relative sentiment relationship of different sentiment vocabulary; according to the relative sentiment relationship of different sentiment vocabulary, obtain the final sentiment degree of each sentiment vocabulary; Add up the final sentiment degrees of all sentiment vocabulary with a final sentiment degree greater than 0 of the a-th medical service information, and record it as the positive sentiment index of the a-th medical service information; add up the final sentiment degrees of all sentiment vocabulary with a final sentiment degree less than 0 of the a-th medical service information, and take the absolute value, and record it as the negative sentiment index of the a-th medical service information.

6. A method for behavior analysis and medical service push based on big data according to claim 5, characterized in that, The obtaining of the sentiment vocabulary set of each medical service information according to the sentiment vocabulary performance of each medical service information includes: Use jieba word segmentation to perform part-of-speech tagging on each medical service information, obtain the words with the part-of-speech of adjective or adverb in each medical service information, and record them as the emotional words of each medical service information; perform union operation on the emotional words of each medical service information to obtain the emotional word set of each medical service information.

7. A method for behavior analysis and medical service push based on big data according to claim 5, characterized in that Based on the voting mechanism, obtain the relative emotional relationship of different emotional words, including: Perform union operation on the emotional words of each medical service information to obtain the emotional word set of each medical service information; Set up a voting mechanism including several personnel. For any two emotional words, each person votes for the emotional word that he / she believes is more positive among the two emotional words. Record the emotional word with the highest number of votes among the two emotional words as the relatively positive emotional word among the two emotional words, and record the emotional word with the lowest number of votes among the two emotional words as the relatively negative emotional word among the two emotional words.

8. A method for behavior analysis and medical service push based on big data according to claim 5, characterized in that According to the relative emotional relationship of different emotional words, obtain the final emotional degree of each emotional word, including: According to the relative emotional relationship of different emotional words, construct an emotional word degree tree, and the emotional word degree tree is a binary sorting tree; perform in-order traversal on the emotional word degree tree to obtain an in-order emotional word sequence; Record the emotional word corresponding to the median of the ordinal numbers of the in-order emotional word sequence as the neutral emotional word, subtract the ordinal number of the neutral emotional word from the ordinal number of the u-th emotional word in the in-order emotional word sequence, and then divide by the maximum ordinal number of the in-order emotional word sequence, and record it as the final emotional degree of the u-th emotional word.

9. A method for behavior analysis and medical service push based on big data according to claim 1, characterized in that, Obtain the credibility of each medical service noun, including: Obtain all medical service information containing the s-th medical service noun, and record it as the relevant medical service information of the s-th medical service noun; ； In the formula, represents the credibility of the s-th medical service term; represents the number of relevant medical service information of the s-th medical service term; represents the credibility of the a-th relevant medical service information of the s-th medical service term; and represent the positive sentiment index and negative sentiment index of the a-th relevant medical service information of the s-th medical service term; represents the absolute value function; represents the exponential function with the natural constant as the base.

10. A behavior analysis and medical service push system based on big data, characterized in that, including: A processor and a memory, the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the method for behavior analysis and medical service push based on big data according to any one of claims 1-9 is implemented.

Images