Health care information push method and system based on social perception technology

Abstract

The present invention provides a health care information push method and system based on social perception technology, the method comprising: obtaining health care comment filter words, and creating an information retriever to crawl comment data according to the health care comment filter words to obtain corresponding health care comment information; performing text processing on the health care comment information to obtain a number of first comment information; performing sentiment analysis on the first comment information to obtain second comment information; obtaining a preset number of the second comment information, combining with a preset reputation value calculation model, to obtain the reputation value of the health care location; generating health care location recommendation information according to the reputation value of the health care location, and pushing the health care location recommendation information to the target user. Therefore, by applying the method described in the present invention, it is possible to effectively promote each health care location to improve its own health care services and assist users in selecting suitable health care locations, and efficiently handle complex health care comments.

Description

Technical Field

[0001] The present invention relates to the field of smart health care, and in particular to a health care information push method and system based on social perception technology. Background Art

[0002] At present, the reputation evaluation of health and wellness locations mainly relies on manual collection and limited user feedback, which has limitations in comprehensiveness and timeliness.

[0003] In addition, there is a wealth of user review data on social media. These data contain important information such as users' emotional attitudes and evaluations of various health and wellness locations. When used to participate in the construction of a health and wellness reputation evaluation system, it will undoubtedly be able to better and more comprehensively demonstrate the advantages and disadvantages of the corresponding health and wellness locations.

[0004] However, existing technical means have shortcomings in processing these user review data, resulting in inefficient information processing, insufficient reliability of obtained information and low promotion efficiency. Even if applied to the health and wellness reputation evaluation of health and wellness locations, it is difficult to obtain users' real feelings and comments in real time and comprehensively. At the same time, it also limits users' ability to make decisions based on sufficient information when faced with numerous health and wellness locations and the efficiency of health and wellness location selection. Summary of the invention

[0005] Based on this, the purpose of the present invention is to provide a health care information push method and system based on social perception technology, which can improve the information processing efficiency of health care review data, the information reliability of pushed health care location recommendation information and the user's health care location selection efficiency.

[0006] The purpose of the present invention can be achieved by the following technical solutions:

[0007] A method for pushing health care information based on social perception technology, comprising the following steps: obtaining health care review filter words, and creating an information retriever to crawl comment data according to the health care review filter words to obtain corresponding health care review information, wherein the health care review filter words include health care locations, and the health care review information is user comment information for the health care locations; performing text processing on the health care review information to obtain a plurality of first comment information, wherein the first comment information includes health care resource evaluation, health care facility evaluation and health care built environment evaluation of the corresponding health care location; performing sentiment analysis on the first comment information to obtain second comment information, wherein the second comment information includes the content in the first comment information sentiment tendency information of health care resource evaluation, sentiment tendency information of health care facility evaluation and sentiment tendency information of health care built environment evaluation; obtaining a preset number of the second comment information, determining the positive feedback rate of the health care resources of the health care location, the positive feedback rate of the health care facilities of the health care location and the positive feedback rate of the health care built environment of the health care location; obtaining the reputation value of the health care location according to the positive feedback rate of the health care resources of the health care location, the positive feedback rate of the health care facilities of the health care location and the positive feedback rate of the health care built environment of the health care location, in combination with a preset reputation value calculation model; generating health care location recommendation information according to the reputation value of the health care location, and pushing the health care location recommendation information to target users.

[0008] A health care information push system based on social perception technology, which includes: a health care comment information acquisition unit, which is used to acquire health care comment filter words, and create an information retriever according to the health care comment filter words to crawl comment data to obtain corresponding health care comment information, wherein the health care comment filter words include health care locations, and the health care comment information is user comment information for the health care locations; a health care comment information text processing unit, which is used to perform text processing on the health care comment information to obtain a plurality of first comment information, wherein the first comment information includes a health care resource evaluation, a health care facility evaluation, and a health care built environment evaluation of the corresponding health care location; a health care comment information sentiment analysis unit, which is used to perform sentiment analysis on the first comment information to obtain second comment information, wherein the second comment information includes the first comment The emotional tendency information of the health care resource evaluation, the emotional tendency information of the health care facility evaluation and the emotional tendency information of the health care built environment evaluation in the information; an emotional tendency calculation unit, used to obtain a preset number of the second comment information, and determine the positive feedback rate of the health care resources of the health care location, the positive feedback rate of the health care facilities of the health care location and the positive feedback rate of the health care built environment of the health care location; a reputation score acquisition unit, used to obtain the reputation value of the health care location according to the positive feedback rate of the health care resources of the health care location, the positive feedback rate of the health care facilities of the health care location and the positive feedback rate of the health care built environment of the health care location, in combination with a preset reputation value calculation model; a health care information push unit, used to generate health care location recommendation information according to the reputation value of the health care location, and push the health care location recommendation information to target users.

[0009] Compared with the prior art, the method described in the present invention first obtains health care review filter words and creates an information retriever, and then crawls the review data according to the health care review filter words to obtain the corresponding health care review information, so that a large number of health care reviews can be preliminarily screened and the information complexity is reduced. Secondly, the obtained health care review information is text processed and classified to obtain the health care resource evaluation, health care facility evaluation and health care built environment evaluation of the corresponding health care location, so as to improve the reliability of the information obtained subsequently. Then, the health care resource evaluation, health care facility evaluation and health care built environment evaluation are respectively subjected to sentiment analysis to obtain the sentiment tendency information of the corresponding evaluation, so that the evaluation information can be analyzed and processed efficiently and accurately. Finally, the positive feedback rate of the health care resources of the health care location, the positive feedback rate of the health care facilities of the health care location and the positive feedback rate of the health care built environment of the health care location are respectively calculated for the sentiment tendency information. At the same time, the reputation value of the health care location is obtained in combination with a preset reputation value calculation model, and the health care location recommendation information is generated, and the health care location recommendation information is pushed to the target user. Therefore, the method described in the present invention can obtain the user's real feelings and comments in real time and comprehensively, allowing the user to make decisions based on sufficient information when faced with numerous health and wellness locations. It also improves the information processing efficiency of health and wellness review data and the promotion efficiency of health and wellness location recommendation information. In addition, combined with the sentiment analysis of the review data, it ensures the information reliability of the pushed health and wellness location recommendation information, thereby improving the user's health and wellness location selection efficiency.

[0010] For better understanding and implementation, the present invention is described in detail below with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] Figure 1 A flowchart of a health care information push method based on social perception technology provided by the present invention;

[0012] Figure 2 A flowchart of the steps of obtaining health care review information in the health care information push method based on social perception technology provided by the present invention;

[0013] Figure 3 A flowchart of the steps of obtaining first comment information in the health care information push method based on social perception technology provided by the present invention;

[0014] Figure 4 A flowchart of the first step of obtaining second comment information in the health care information push method based on social perception technology provided by the present invention;

[0015] Figure 5 A flowchart of the second step of obtaining the second comment information in the health care information push method based on social perception technology provided by the present invention;

[0016] Figure 6 A schematic diagram of a health care information push system based on social perception technology provided by the present invention. DETAILED DESCRIPTION

[0017] The present invention provides a health care information push method and system based on social perception technology. In order to make the purpose, technical solution and effect of the present invention clearer and more specific, the present invention is further described in detail with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0018] It will be understood by those skilled in the art that, unless expressly stated, the singular forms "one", "said", and "the" used herein may also include plural forms. It should be further understood that the term "comprising" used in the specification of the present invention refers to the presence of the features, integers, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or groups thereof. It should be understood that when we refer to an element as being "connected" or "coupled" to another element, it may be directly connected or coupled to the other element, or there may be intermediate elements. In addition, the "connection" or "coupling" used herein may include wireless connection or wireless coupling. The term "and / or" used herein includes all or any unit and all combinations of one or more associated listed items.

[0019] It will be understood by those skilled in the art that, unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as generally understood by those skilled in the art in the art to which the present invention belongs. It should also be understood that terms such as those defined in general dictionaries should be understood to have meanings consistent with the meanings in the context of the prior art, and will not be interpreted with idealized or overly formal meanings unless specifically defined as herein.

[0020] The invention is further described below by describing embodiments in conjunction with the accompanying drawings.

[0021] Example 1

[0022] Please refer to Figure 1 , please refer to Figure 1 , Figure 1 A flowchart of a health care information push method based on social perception technology provided by the present invention. The method comprises the following steps:

[0023] S10, obtaining a health care review filter word, and creating an information retriever to crawl review data according to the health care review filter word to obtain corresponding health care review information, wherein the health care review filter word includes a health care location, and the health care review information is user review information for the health care location;

[0024] S20, performing text processing on the health care review information to obtain a plurality of first review information, wherein the first review information includes a health care resource evaluation, a health care facility evaluation, and a health care built environment evaluation of a corresponding health care location;

[0025] S30, performing sentiment analysis on the first comment information to obtain second comment information, wherein the second comment information includes sentiment tendency information of health care resource evaluation, sentiment tendency information of health care facility evaluation, and sentiment tendency information of health care built environment evaluation in the first comment information;

[0026] S40, obtaining a preset number of pieces of second review information, and determining a positive feedback rate of health care resources of the health care location, a positive feedback rate of health care facilities of the health care location, and a positive feedback rate of health care built environment of the health care location;

[0027] S50, obtaining the reputation value of the health care location according to the positive feedback rate of the health care resources of the health care location, the positive feedback rate of the health care facilities of the health care location, and the positive feedback rate of the health care built environment of the health care location, in combination with a preset reputation value calculation model;

[0028] S60: Generate health care location recommendation information according to the reputation value of the health care location, and push the health care location recommendation information to the target user.

[0029] For step S10, obtain health care review filter words, and create an information retriever to crawl review data based on the health care review filter words to obtain corresponding health care review information;

[0030] Among them, the health care review filter words refer to keywords input by the user for the information retriever to filter health care review information. In one embodiment, the health care review filter words include health care locations (such as Bama and Lijiang); the health care review information is user review information for the health care locations; the information retriever is a retrieval tool that formulates retrieval strategies based on the user's information needs to conduct targeted information searches. In one embodiment, the information retriever is a web crawler.

[0031] In one embodiment, the health review filter words also include a target retrieval time period, a target social media platform, and a target retrieval cycle; the target retrieval time period refers to the expected time period selected and set by the user for capturing the corresponding health review information. In one embodiment, the target retrieval time period may be within the last month or within half a year. When the target retrieval time period set by the user is within the last month, the information retriever will obtain and filter the health review information within the last month; the target social media platform refers to a content generation and exchange platform based on user relationships on the Internet. In one embodiment, the target social media platform may be Weibo, Tieba, or Zhihu; the target retrieval cycle refers to the comment data capture cycle for the corresponding health review information selected and set by the user. In one embodiment, the target retrieval cycle may be once a day or once a week. When the target retrieval cycle set by the user is once a day, the information retriever will perform comment data capture for the corresponding health review information once a day.

[0032] Please refer to Figure 2 In one embodiment, the step S10 specifically includes the following steps:

[0033] S101. Create an information retriever in response to a health care location, a target search time period, and a target social media platform input by a user.

[0034] S102, controlling the information retriever to determine whether the target social media platform has an API interface, and if so, controlling the information retriever to obtain the API interface of the target social media platform, and obtaining corresponding health care review information through the API interface according to the health care location and the target retrieval time period.

[0035] Among them, the API interface refers to a set of predefined methods and protocols that enable different software systems to communicate with each other and share data by defining a unified data format and protocol. In one embodiment, the API interface refers to a data interaction interface for performing information transmission.

[0036] In one embodiment, the information retriever is controlled to access and obtain the API interface of Weibo, and then through the API interface of Weibo, the structured data of the corresponding health care review information is obtained and downloaded according to the health care location and the target retrieval time period. This can shorten the time the user waits to obtain data information, and effectively improve the information acquisition rate of the information retriever in obtaining health care review information from Weibo.

[0037] S103. If not, control the information retriever to obtain the uniform resource locator of the target social media platform and locate the search component of the target social media platform; according to the uniform resource locator of the target social media platform, send the health care location to the search component of the target social media platform, obtain the search result returned by the search component of the target social media platform, and control the information retriever to obtain the health care review information in the search result by sending a hypertext transfer protocol request.

[0038] In one embodiment, if the information retriever fails to obtain the health care review information in the search results by sending a hypertext transfer protocol request, it is determined whether the generation of the hypertext transfer protocol request fails. If so, it is retried to generate the hypertext transfer protocol request a preset number of times. If not, it is determined whether the target social media platform implements an IP address restriction policy. If so, the information retriever is controlled to send the hypertext transfer protocol request through a proxy IP address.

[0039] S104. According to the target search time period, determine whether the release time of the health care review information in the search results is within the target search time period. If so, perform content deduplication and format cleaning to obtain the corresponding health care review information.

[0040] Among them, the content deduplication refers to filtering and deleting duplicate comments or data in the health care review information in the search results; the format cleaning refers to removing HTML tags and special characters in the health care review information in the search results.

[0041] In one embodiment, the obtained health care review information is locally stored through a MySQL database, wherein the data table structure of the MySQL database is as shown in Table 1.

[0042] Table 1 Schematic diagram of the data table structure of MySQL database

[0043]

[0044] In step S20, text processing is performed on the health care review information to obtain a plurality of first review information;

[0045] Among them, the first review information includes the health care resource evaluation, health care facility evaluation and health care built environment evaluation of the corresponding health care location.

[0046] Please refer to Figure 3 In one embodiment, the step S20 specifically includes the following steps:

[0047] S201. Call a natural language processing tool to perform data cleaning on the health care review information according to preset data cleaning rules.

[0048] The data cleaning includes at least removing punctuation marks, removing stop words and removing web page tags.

[0049] In one embodiment, the natural language processing tool may be an nltk library or a spaCy library.

[0050] S202. Calling a natural language processing tool to perform semantic recognition on the health care review information after data cleaning, and performing sentence segmentation on the health care review information after data cleaning according to a preset text segmentation rule to obtain a plurality of first review information.

[0051] In one embodiment, the health care review information after data cleaning is segmented by calling a statistical-based word segmenter (jieba). Specifically, for example, for a Weibo comment "This forest is very beautiful, the air is fresh, and the negative oxygen ions are high. The scenery beside the lake is also good, but I think the forest here is a bit deserted", after sentence segmentation, the Weibo comment will be split into the following words: ['this piece', 'forest', 'very', 'beautiful', 'air', 'fresh', 'negative oxygen ions', 'very', 'high', 'lake', 'beside', 'scenery', 'nice', 'but', 'I', 'feel', 'here', 'forest', 'somewhat', 'deserted'].

[0052] In step S30, sentiment analysis is performed on the first comment information to obtain second comment information;

[0053] Among them, the second comment information includes the sentiment tendency information of the health care resource evaluation, the sentiment tendency information of the health care facility evaluation and the sentiment tendency information of the health care built environment evaluation in the first comment information.

[0054] Please refer to Figure 4 In one embodiment, the step S30 specifically includes the following steps:

[0055] S301, obtaining a training set of health care review texts, and according to a priori probability calculation formula, counting and respectively calculating the priori probabilities that the texts in the training set of health care review texts belong to positive emotions and negative emotions:

[0056] ,

[0057] in, P (pos) is the prior probability that the text in the health review text training set belongs to positive emotion, P(neg) is the prior probability that the text in the health review text training set belongs to negative sentiment, N is the number of texts in the health review text training set, N pos is the number of texts with positive emotions in the health review text training set ，N neg is the number of texts with negative sentiment in the health review text training set.

[0058] In one embodiment, the health review text training set includes a number of first review information with known emotional tendencies; the first review information in the health review text training set is stored in the form of text; and the emotional tendencies include positive and negative.

[0059] S302, according to the preset evaluation vocabulary and conditional probability calculation formula, in combination with the texts in the health review text training set, respectively calculate the probability that the words in the evaluation vocabulary appear in the texts belonging to positive emotions in the health review text training set and the probability that the words in the health review text training set belong to negative emotions:

[0060] ,

[0061] in, P ( x │pos) is a word in the evaluation vocabulary x The probability of text with positive sentiment appearing in the health review text training set, P ( x │neg) The words in the evaluation vocabulary x The probability of text with negative sentiment appearing in the health review text training set, count( x ,pos) is a word in the evaluation vocabulary x The number of times it appears in the text with positive emotions in the health review text training set, count( x ,neg) is a word in the evaluation vocabulary x The number of times it appears in the text with negative sentiment in the health review text training set, a is the Laplace smoothing parameter, │ V │ is the size of the evaluation vocabulary, total(pos) is the total number of all words in the evaluation vocabulary that appear in the texts with positive emotions in the health review text training set, and total(neg) is the total number of all words in the evaluation vocabulary that appear in the texts with negative emotions in the health review text training set.

[0062] In one embodiment, the evaluation vocabulary includes the vocabulary in the first comment information and the vocabulary of the first comment information in the health review text training set; the value of the Laplace smoothing parameter is 1.

[0063] S303. According to the posterior probability calculation formula, combined with the prior probabilities that the texts in the health review text training set belong to positive emotions and negative emotions, and the probabilities that the words in the evaluation vocabulary appear in the texts in the health review text training set that belong to positive emotions, and the probabilities that the texts in the health review text training set that belong to negative emotions, a naive Bayes classifier is constructed and trained:

[0064] ,

[0065] in, P ( c │ y ) is the given text y belong c The probability of class emotion, P ( y │ c ) is the given text y The texts appearing in the health review text training set belong to c The probability of sentiment-like text, P( c ) is the text in the health review text training set belonging to c The prior probability of the class emotion, c Emotions include positive emotions and negative emotions. P ( y ) is a preset constant; and, wherein the given text y Depend on n The vocabulary in the evaluation vocabulary is composed of:

[0066] ,

[0067] in, y i For the y i The words mentioned, P ( y i │ c ) is the vocabulary in the given text y y i The texts appearing in the health review text training set belong to c The probability of text belonging to the class sentiment.

[0068] In one embodiment, by y 1 Input into the naive Bayes classifier to obtain the given text y1 of P (pos│ y 1) and P (neg│ y 1) By judging the sizes of the two, the larger one is obtained as the given text y The classification result of 1 (i.e. sentiment tendency) is

[0069] In one embodiment, by inputting several first comment information with known sentiment tendencies into the Naive Bayes classifier respectively, it is determined whether the obtained classification results are the same as their sentiment tendencies, thereby determining whether the Naive Bayes classifier has completed training.

[0070] S304: taking the health care resource evaluation, health care facility evaluation and health care built environment evaluation of the health care location in the first comment information as the given text y Input to the naive Bayes classifier that has completed training.

[0071] S305: Obtain the probability of the health care resource evaluation, the health care facility evaluation, and the health care built environment evaluation each belonging to a positive emotion and a negative emotion.

[0072] S306. Compare the probabilities of positive emotions and negative emotions of the health care resource evaluation, the health care facility evaluation, and the health care built environment evaluation, respectively, obtain the emotional tendency information of the health care resource evaluation, the emotional tendency information of the health care facility evaluation, and the emotional tendency information of the health care built environment evaluation, and obtain the second comment information.

[0073] The emotional tendency information includes positive and negative.

[0074] In one embodiment, if the probability that the health care built environment evaluation belongs to positive emotion is greater than the probability that it belongs to negative emotion, then the classification result of the health care built environment evaluation in the naive Bayes classifier is determined to be positive emotion, that is, the emotional tendency information of the health care built environment evaluation is positive; otherwise it is negative; the same applies to the health care resource evaluation and the health care facility evaluation.

[0075] In one embodiment, for example,

[0076] Input: Sentence: This forest is very beautiful, the air is fresh and the negative oxygen ions are high.

[0077] Output: Sentiment: positive ( P (pos│ y 1): P (neg│ y1) =0.89:0.11). —— Front

[0078] Input: Sentence: The scenery by the lake is also nice, but I think the forest here is a bit deserted.

[0079] Output: Sentiment: negative ( P (pos│ y 1): P (neg│ y 1) =0.23:0.77). ——Negative

[0080] In addition, the health care information push method based on social perception technology of the present invention also provides some health care information push steps different from steps S301-S306, which are used to perform sentiment analysis on the first comment information in different ways to obtain the second comment information. Please refer to Figure 5 , the steps are as follows:

[0081] S311, assigning scores to the words in the first comment information according to a preset word sentiment score table, and calculating the sentiment score of the health care resource evaluation of the health care location, the sentiment score of the health care facility evaluation, and the sentiment score of the health care built environment evaluation in the first comment information;

[0082] S312. According to a preset sentiment tendency prediction criterion, combined with the sentiment scores of the health care resource evaluation, the health care facility evaluation and the health care built environment evaluation, the sentiment tendency information of the health care resource evaluation, the sentiment tendency information of the health care facility evaluation and the sentiment tendency information of the health care built environment evaluation are obtained to obtain the second comment information.

[0083] In one embodiment, step S311 specifically includes the following steps:

[0084] S311a: Perform part-of-speech judgment on the words in the first comment information to obtain adjectives, nouns and adverbs in the first comment information.

[0085] Among them, commonly used adjectives include comfortable, safe and quiet; commonly used nouns include forest, river and hospital; commonly used adverbs include very, extremely and very.

[0086] In one embodiment, the part of speech of the words in the first comment information is determined by applying a natural language processing tool.

[0087] S311b. Assign scores to the adjectives, nouns and adverbs in the first comment information according to a preset vocabulary sentiment score table.

[0088] In one embodiment, the vocabulary sentiment score table can be determined by constructing a deep learning model, or by an empirical evaluation scoring mechanism.

[0089] S311c, calculating the sentiment scores of different evaluations in the first comment information according to the sentiment score calculation formula:

[0090] ,

[0091] in, G is the sentiment score of the evaluation in the first comment information, n is the number of adjective-noun-adverb word combinations in the evaluation in the first review information, Gi (adv) is the first comment in the evaluation of the first comment information. i Scoring of adverbs in adjective-noun-adverb combinations, Gi (adj) is the first comment in the evaluation of the first comment information. i Scoring of adjectives in adjective-noun-adverb combinations, Gi (n) is the first comment in the evaluation of the first comment information. i Scoring of nouns in lexical combinations of adjective-noun-adverb.

[0092] Among them, the vocabulary combination includes adjectives, nouns and adverbs; the adjectives, nouns and adverbs constitute a combination of adjective-noun-adverb. In some other alternative implementations, the three words can also be interchanged in order in the same vocabulary combination to form different combination forms to continue to implement the remaining steps of the method described in the present invention; if a certain word does not exist in the vocabulary combination, the score of this word in the vocabulary combination is set to 1.

[0093] In one embodiment, the health care resource evaluation in the first comment information, the health care facility evaluation in the first comment information, and the health care built environment evaluation in the first comment information respectively include a plurality of adjective-noun-adverb vocabulary combinations.

[0094] Specifically, according to the sentiment score calculation formula, the sentiment score of a certain evaluation can be obtained by calculating the sentiment scores of each word combination in the evaluation and then accumulating them. For example, for a microblog comment "This forest is very beautiful, the air is fresh, and the negative oxygen ion content is very high, but I think the forest here is a bit deserted.", the sentiment score of the microblog comment is: 1 (forest) * 2 (very) * 2 (beautiful) + 1 (air) * 2 (fresh) * 1 + 1 (negative oxygen ions) * 1.5 (very) * 2 (high) + 1 (forest) * -1 (deserted) = 8. Combined with the preset sentiment tendency prediction criteria in step S312, the sentiment tendency information of the comment is determined to be positive.

[0095] S311d. Obtain the sentiment score of the health care resource evaluation, the sentiment score of the health care facility evaluation, and the sentiment score of the health care built environment evaluation of the health care location in the first comment information.

[0096] In one embodiment, steps S311a-S311c can be combined with relevant steps of training the naive Bayes classifier, that is, combined with steps S301-S303, so as to improve the classification accuracy of the trained naive Bayes classifier for each evaluation, and thereby obtain more accurate sentiment tendency information of each evaluation.

[0097] For step S40, a preset number of the second review information is obtained, and the positive feedback rate of the health care resources of the health care location, the positive feedback rate of the health care facilities of the health care location, and the positive feedback rate of the health care built environment of the health care location are determined;

[0098] In one embodiment, step S40 specifically includes the following steps:

[0099] S401: Obtain a preset number of pieces of the second comment information and construct an evaluation information set.

[0100] S402, dividing the number of health care resource evaluations with a positive sentiment tendency in the evaluation information set by the number of health care resource evaluations in the evaluation information set, dividing the number of health care facility evaluations with a positive sentiment tendency in the evaluation information set by the number of health care facility evaluations in the evaluation information set, and dividing the number of health care built environment evaluations with a positive sentiment tendency in the evaluation information set by the number of health care built environment evaluations in the evaluation information set;

[0101] S403, determining the positive feedback rate of the health care resources of the health care location, the positive feedback rate of the health care facilities of the health care location, and the positive feedback rate of the health care built environment of the health care location.

[0102] For step S50, according to the positive feedback rate of the health care resources of the health care location, the positive feedback rate of the health care facilities of the health care location, and the positive feedback rate of the health care built environment of the health care location, in combination with a preset reputation value calculation model, the reputation value of the health care location is obtained;

[0103] In one embodiment, step S50 specifically includes the following steps:

[0104] S501. In combination with preset standardized score conversion criteria, the positive feedback rate of the health and wellness resources of the health and wellness location, the positive feedback rate of the health and wellness facilities of the health and wellness location, and the positive feedback rate of the health and wellness built environment of the health and wellness location are converted into the health and wellness resource score of the health and wellness location, the health and wellness facility score of the health and wellness location, and the health and wellness built environment score of the health and wellness location, respectively.

[0105] In one embodiment, if the positive feedback rate is >0.8, the score is 90-100 points; if the positive feedback rate is >0.6, the score is 70-90 points; if the positive feedback rate is >0.4, the score is 50-70 points; if the positive feedback rate is <0.4, the score is below 50 points.

[0106] S502. According to the reputation value calculation formula, the health care resource score of the health care location, the health care facility score of the health care location, and the health care built environment score of the health care location are multiplied by corresponding preset weights and then added together to obtain the reputation value of the health care location:

[0107] ,

[0108] in, V (rep) is the reputation value of the health care location. n =3, w i The preset weights corresponding to health care resources, health care facilities or health care built environment, p i It is the health care resource score of the health care location, the health care facility score of the health care location, or the health care built environment score of the health care location.

[0109] In step S60, according to the reputation value of the health care location, health care location recommendation information is generated, and the health care location recommendation information is pushed to the target user;

[0110] Among them, the health and wellness location recommendation information at least includes the recommendation level of the health and wellness location, the health and wellness resource score of the health and wellness location, the health and wellness facility score of the health and wellness location, the health and wellness built environment score of the health and wellness location and the health and wellness review information of the health and wellness location.

[0111] In one embodiment, when one of the health care resource score of the health care location, the health care facility score of the health care location, and the health care built environment score of the health care location is less than 50 points, it is color-coded and a warning message is output.

[0112] Example 2

[0113] Please refer to Figure 6 The present invention also provides a health care information push system based on social perception technology to implement the steps of the health care information push method based on social perception technology described in the above embodiment. The health care information push system based on social perception technology includes: a health care review information acquisition unit 301, a health care review information text processing unit 302, a health care review information sentiment analysis unit 303, a sentiment tendency calculation unit 304, a reputation score acquisition unit 305 and a health care information push unit 306.

[0114] The health care review information acquisition unit 301 is used to acquire health care review filter words, and create an information retriever to crawl review data according to the health care review filter words to obtain corresponding health care review information, wherein the health care review filter words include health care locations, and the health care review information is user review information for the health care locations;

[0115] The health care review information text processing unit 302 is used to perform text processing on the health care review information to obtain a plurality of first review information, wherein the first review information includes a health care resource evaluation, a health care facility evaluation, and a health care built environment evaluation of a corresponding health care location;

[0116] The health care review information sentiment analysis unit 303 is used to perform sentiment analysis on the first review information to obtain second review information, wherein the second review information includes sentiment tendency information of health care resource evaluation, sentiment tendency information of health care facility evaluation and sentiment tendency information of health care built environment evaluation in the first review information;

[0117] The sentiment tendency calculation unit 304 is used to obtain a preset number of the second review information, and determine the positive feedback rate of the health care resources of the health care location, the positive feedback rate of the health care facilities of the health care location, and the positive feedback rate of the health care built environment of the health care location;

[0118] The reputation score acquisition unit 305 is used to acquire the reputation value of the health care location according to the positive feedback rate of the health care resources of the health care location, the positive feedback rate of the health care facilities of the health care location, and the positive feedback rate of the health care built environment of the health care location, in combination with a preset reputation value calculation model;

[0119] The health care information pushing unit 306 is used to generate health care location recommendation information according to the reputation value of the health care location, and push the health care location recommendation information to the target user.

[0120] It should be noted that the above-mentioned embodiment provides a health care information push system based on social perception technology. When implementing a health care information push method based on social perception technology, it only uses the division of the above-mentioned functional modules as an example. In actual applications, the above-mentioned functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above.

[0121] In addition, the health care information push system based on social perception technology provided in the above embodiment and the health care information push method based on social perception technology in Example 1 belong to the same concept. The implementation process thereof is detailed in the method embodiment, i.e., Example 1, and will not be repeated here.

[0122] Compared with the prior art, the method described in the present invention can scientifically measure the reputation of health and wellness tourism areas and comprehensively analyze its influencing factors, which is conducive to more accurately grasping the sustainable development capabilities of the region's health and wellness industry. It is of great significance to solidly promote the rational development, quality improvement, efficiency enhancement and transformation development of health and wellness tourism. While conducting a comprehensive and dynamic evaluation of the reputation of health and wellness institutions, it provides consumers with more reliable information and provides a basis for improvement for the health and wellness institutions themselves at various health and wellness locations.

[0123] The above-mentioned embodiments only express several implementation methods of the present invention, and the description is relatively specific and detailed, but it cannot be understood as limiting the scope of the invention patent. It should be pointed out that for ordinary technicians in this field, several modifications and improvements can be made without departing from the concept of the present invention, and the present invention is also intended to include these modifications and modifications.

Claims

1. A method for pushing health care information based on social perception technology, comprising the following steps: Acquire a health care review filter word, and create an information retriever to crawl review data according to the health care review filter word to obtain corresponding health care review information, wherein the health care review filter word includes a health care location, and the health care review information is user review information for the health care location; Performing text processing on the health care review information to obtain a plurality of first review information, wherein the first review information includes a health care resource evaluation, a health care facility evaluation, and a health care built environment evaluation of a corresponding health care location; Perform sentiment analysis on the first comment information to obtain second comment information, wherein the second comment information includes sentiment tendency information of health care resource evaluation, sentiment tendency information of health care facility evaluation and sentiment tendency information of health care built environment evaluation in the first comment information, specifically: construct and train a naive Bayes classifier, including: obtaining a health care comment text training set, and according to a priori probability calculation formula, statistically and respectively calculating the prior probabilities that the texts in the health care comment text training set belong to positive emotions and negative emotions: ， in, P (pos) is the prior probability that the text in the health review text training set belongs to positive emotion, P (neg) is the prior probability that the text in the health review text training set belongs to negative sentiment, N is the number of texts in the health review text training set, N pos is the number of texts with positive emotions in the health review text training set ，N neg is the number of texts with negative sentiment in the health review text training set; According to the preset evaluation vocabulary and conditional probability calculation formula, combined with the texts in the health review text training set, the probability that the words in the evaluation vocabulary appear in the texts with positive emotions in the health review text training set and the probability that the texts with negative emotions in the health review text training set are calculated respectively: ， in, P ( x │pos) is a word in the evaluation vocabulary x The probability of text with positive sentiment appearing in the health review text training set, P ( x │neg) The words in the evaluation vocabulary x The probability of text with negative sentiment appearing in the health review text training set, count( x ,pos) is a word in the evaluation vocabulary x The number of times it appears in the text with positive emotions in the health review text training set, count( x ,neg) is a word in the evaluation vocabulary x The number of times it appears in the text with negative sentiment in the health review text training set, a is the Laplace smoothing parameter, │ V │ is the size of the evaluation vocabulary, total(pos) is the total number of all words in the evaluation vocabulary that appear in the texts with positive emotions in the health review text training set, and total(neg) is the total number of all words in the evaluation vocabulary that appear in the texts with negative emotions in the health review text training set; According to the posterior probability calculation formula, combined with the prior probabilities that the texts in the health review text training set belong to positive emotions and negative emotions, and the probabilities that the words in the evaluation vocabulary appear in the texts belonging to positive emotions in the health review text training set and the probabilities that the texts in the health review text training set belong to negative emotions, a naive Bayes classifier is constructed and trained: ， Among them, P( c │ y ) is the given text y belong c The probability of class sentiment, P( y │ c ) is the given text y The texts appearing in the health review text training set belong to c The probability of sentiment-like text, P( c ) is the text in the health review text training set belonging to c The prior probability of the class emotion, c Class emotions include positive emotions and negative emotions, P( y ) is a preset constant; And, the given text y Depend on n The vocabulary in the evaluation vocabulary is composed of: ， in, y i For the y i The words, P( y i │ c ) is the vocabulary in the given text y y i The texts appearing in the health review text training set belong to c The probability of sentiment-like text; Inputting the health care resource evaluation, health care facility evaluation and health care built environment evaluation of the health care location in the first review information as given texts into the trained naive Bayes classifier respectively; Obtaining the probability of each of the health care resource evaluation, the health care facility evaluation, and the health care built environment evaluation belonging to a positive emotion and the probability of each of the health care resource evaluation, the health care facility evaluation, and the health care built environment evaluation belonging to a negative emotion; Respectively comparing the probability of positive emotion and the probability of negative emotion of the health care resource evaluation, the health care facility evaluation, and the health care built environment evaluation, obtaining the emotional tendency information of the health care resource evaluation, the emotional tendency information of the health care facility evaluation, and the emotional tendency information of the health care built environment evaluation, and obtaining the second comment information; Obtaining a preset number of pieces of the second review information, and determining a positive feedback rate of the health care resources of the health care location, a positive feedback rate of the health care facilities of the health care location, and a positive feedback rate of the health care built environment of the health care location; According to the positive feedback rate of the health care resources of the health care location, the positive feedback rate of the health care facilities of the health care location, and the positive feedback rate of the health care built environment of the health care location, in combination with a preset reputation value calculation model, the reputation value of the health care location is obtained; According to the reputation value of the health and wellness location, health and wellness location recommendation information is generated, and the health and wellness location recommendation information is pushed to the target user.

2. The health care information push method based on social perception technology according to claim 1 is characterized by: The health review filter words also include a target search time period, a target social media platform, and a target search cycle; The steps of obtaining health care review filter words, and creating an information retriever to crawl review data according to the health care review filter words to obtain corresponding health care review information include: Creating an information retriever in response to a health care location, a target search time period, and a target social media platform input by a user; Control the information retriever to determine whether the target social media platform has an API interface, and if so, control the information retriever to obtain the API interface of the target social media platform, and obtain corresponding health care review information through the API interface according to the health care location and the target search time period; If not, control the information retriever to obtain the uniform resource locator of the target social media platform and locate the search component of the target social media platform, send the health care location to the search component of the target social media platform according to the uniform resource locator of the target social media platform, obtain the search result returned by the search component of the target social media platform, and control the information retriever to obtain the health care review information in the search result by sending a hypertext transfer protocol request; According to the target search time period, determine whether the release time of the health care review information in the search results is within the target search time period. If so, perform content deduplication and format cleaning to obtain the corresponding health care review information.

3. According to the method for pushing health care information based on social perception technology as described in claim 1, the step of performing text processing on the health care comment information to obtain a plurality of first comment information comprises: Calling a natural language processing tool to perform data cleaning on the health care review information according to a preset data cleaning rule, wherein the data cleaning at least includes removing punctuation marks, removing stop words, and removing web page tags; A natural language processing tool is called to perform semantic recognition on the health care review information after data cleaning, and sentence segmentation is performed on the health care review information after data cleaning according to a preset text segmentation rule to obtain a plurality of first review information.

4. According to the health care information push method based on social perception technology in claim 1, the step of performing sentiment analysis on the first comment information to obtain the second comment information comprises: According to a preset vocabulary sentiment score table, the vocabulary in the first comment information is scored to calculate the sentiment score of the health care resource evaluation of the health care location, the sentiment score of the health care facility evaluation, and the sentiment score of the health care built environment evaluation in the first comment information; According to the preset sentiment tendency prediction criteria, the sentiment tendency information of the health care resources evaluation, the health care facilities evaluation and the health care built environment evaluation are combined with their respective sentiment scores to obtain the sentiment tendency information of the health care resources evaluation, the sentiment tendency information of the health care facilities evaluation and the sentiment tendency information of the health care built environment evaluation to obtain the second comment information.

5. According to the method for pushing health care information based on social perception technology in claim 4, the steps of assigning scores to the words in the first comment information according to a preset word sentiment score table, and calculating the sentiment score of the health care resource evaluation of the health care location in the first comment information, the sentiment score of the health care facility evaluation, and the sentiment score of the health care built environment evaluation include: Performing part-of-speech judgment on the words in the first comment information to obtain adjectives, nouns and adverbs in the first comment information; According to a preset vocabulary sentiment score table, scoring is performed on the adjectives, nouns and adverbs in the first comment information; According to the sentiment score calculation formula, the sentiment scores of different evaluations in the first comment information are calculated: ， in, G is the sentiment score of the evaluation in the first comment information, n is the number of adjective-noun-adverb word combinations in the evaluation in the first review information, Gi (adv) is the first comment in the evaluation of the first comment information. i Scoring of adverbs in adjective-noun-adverb combinations, Gi (adj) is the first comment in the evaluation of the first comment information. i Scoring of adjectives in adjective-noun-adverb combinations, Gi (n) is the first comment in the evaluation of the first comment information. i The assignment of points to nouns in the lexical combination of adjective-noun-adverb; The sentiment score of the health care resource evaluation, the sentiment score of the health care facility evaluation, and the sentiment score of the health care built environment evaluation of the health care location in the first comment information are obtained.

6. According to the method for pushing health care information based on social perception technology in claim 1, the steps of obtaining a preset number of the second review information and determining the positive feedback rate of the health care resources of the health care location, the positive feedback rate of the health care facilities of the health care location, and the positive feedback rate of the health care built environment of the health care location include: Obtaining a preset number of pieces of the second comment information to construct an evaluation information set; Divide the number of health care resource evaluations with a positive sentiment tendency in the evaluation information set by the number of health care resource evaluations in the evaluation information set, divide the number of health care facility evaluations with a positive sentiment tendency in the evaluation information set by the number of health care facility evaluations in the evaluation information set, and divide the number of health care built environment evaluations with a positive sentiment tendency in the evaluation information set by the number of health care built environment evaluations in the evaluation information set; Determine the positive feedback rate of the health and wellness resources of the health and wellness location, the positive feedback rate of the health and wellness facilities of the health and wellness location, and the positive feedback rate of the health and wellness built environment of the health and wellness location.

7. According to the method for pushing health care information based on social perception technology in claim 1, according to the positive feedback rate of health care resources of the health care location, the positive feedback rate of health care facilities of the health care location, and the positive feedback rate of health care built environment of the health care location, combined with a preset reputation value calculation model, the step of obtaining the reputation value of the health care location comprises: In combination with the preset standardized score conversion criteria, the positive feedback rate of the health care resources of the health care location, the positive feedback rate of the health care facilities of the health care location, and the positive feedback rate of the health care built environment of the health care location are converted into the health care resource score of the health care location, the health care facility score of the health care location, and the health care built environment score of the health care location, respectively; According to the reputation value calculation formula, the health care resource score of the health care location, the health care facility score of the health care location, and the health care built environment score of the health care location are multiplied by the corresponding preset weights and then added together to obtain the reputation value of the health care location: ， in, V (rep) is the reputation value of the health care location. n =3, w i The preset weights corresponding to health care resources, health care facilities or health care built environment, p i It is the health care resource score of the health care location, the health care facility score of the health care location, or the health care built environment score of the health care location.

8. A health care information push system based on social perception technology, characterized in that: The health care information push system based on social perception technology includes: A health care review information acquisition unit is used to acquire health care review filter words, and create an information retriever to crawl review data according to the health care review filter words to obtain corresponding health care review information, wherein the health care review filter words include health care locations, and the health care review information is user review information for the health care locations; A health care review information text processing unit, configured to perform text processing on the health care review information to obtain a plurality of first review information, wherein the first review information includes a health care resource evaluation, a health care facility evaluation, and a health care built environment evaluation of a corresponding health care location; The health care review information sentiment analysis unit is used to perform sentiment analysis on the first review information to obtain second review information, wherein the second review information includes the sentiment tendency information of the health care resource evaluation, the sentiment tendency information of the health care facility evaluation and the sentiment tendency information of the health care built environment evaluation in the first review information, specifically: constructing and training a naive Bayes classifier, including: obtaining a health care review text training set, and according to a priori probability calculation formula, statistically calculating and respectively calculating the priori probability that the text in the health care review text training set belongs to positive emotion and negative emotion: ， in, P (pos) is the prior probability that the text in the health review text training set belongs to positive emotion, P (neg) is the prior probability that the text in the health review text training set belongs to negative sentiment, N is the number of texts in the health review text training set, N pos is the number of texts with positive emotions in the health review text training set ，N neg is the number of texts with negative sentiment in the health review text training set; According to the preset evaluation vocabulary and conditional probability calculation formula, combined with the texts in the health review text training set, the probability that the words in the evaluation vocabulary appear in the texts with positive emotions in the health review text training set and the probability that the texts with negative emotions in the health review text training set are calculated respectively: ， in, P ( x │pos) is a word in the evaluation vocabulary x The probability of text with positive sentiment appearing in the health review text training set, P ( x │neg) The words in the evaluation vocabulary x The probability of text with negative sentiment appearing in the health review text training set, count( x ,pos) is a word in the evaluation vocabulary x The number of times it appears in the text with positive emotions in the health review text training set, count( x ,neg) is a word in the evaluation vocabulary x The number of times it appears in the text with negative sentiment in the health review text training set, a is the Laplace smoothing parameter, │ V │ is the size of the evaluation vocabulary, total(pos) is the total number of all words in the evaluation vocabulary that appear in the texts with positive emotions in the health review text training set, and total(neg) is the total number of all words in the evaluation vocabulary that appear in the texts with negative emotions in the health review text training set; According to the posterior probability calculation formula, combined with the prior probabilities that the texts in the health review text training set belong to positive emotions and negative emotions, and the probabilities that the words in the evaluation vocabulary appear in the texts belonging to positive emotions in the health review text training set and the probabilities that the texts in the health review text training set belong to negative emotions, a naive Bayes classifier is constructed and trained: ， Among them, P( c │ y ) is the given text y belong c The probability of class sentiment, P( y │ c ) is the given text y The texts appearing in the health review text training set belong to c The probability of sentiment-like text, P( c ) is the text in the health review text training set belonging to c The prior probability of the class emotion, c Class emotions include positive emotions and negative emotions, P( y ) is a preset constant; And, the given text y Depend on n The vocabulary in the evaluation vocabulary is composed of: ， in, y i For the y i The words, P( y i │ c ) is the vocabulary in the given text y y i The texts appearing in the health review text training set belong to c The probability of sentiment-like text; Inputting the health care resource evaluation, health care facility evaluation and health care built environment evaluation of the health care location in the first review information as given texts into the trained naive Bayes classifier respectively; Obtaining the probability of each of the health care resource evaluation, the health care facility evaluation, and the health care built environment evaluation belonging to a positive emotion and the probability of each of the health care resource evaluation, the health care facility evaluation, and the health care built environment evaluation belonging to a negative emotion; Respectively comparing the probability of positive emotion and the probability of negative emotion of the health care resource evaluation, the health care facility evaluation, and the health care built environment evaluation, obtaining the emotional tendency information of the health care resource evaluation, the emotional tendency information of the health care facility evaluation, and the emotional tendency information of the health care built environment evaluation, and obtaining the second comment information; an emotion tendency calculation unit, configured to obtain a preset number of pieces of the second comment information, and determine a positive feedback rate of the health care resources of the health care location, a positive feedback rate of the health care facilities of the health care location, and a positive feedback rate of the health care built environment of the health care location; A reputation score acquisition unit, configured to acquire the reputation value of the health care location according to the positive feedback rate of the health care resources of the health care location, the positive feedback rate of the health care facilities of the health care location, and the positive feedback rate of the health care built environment of the health care location, in combination with a preset reputation value calculation model; The health care information push unit is used to generate health care location recommendation information according to the reputation value of the health care location, and push the health care location recommendation information to the target user.

Images