Anesthesia protocol generation system based on big data analysis

Abstract

The application discloses an anesthetic scheme generation system based on big data analysis and relates to the field of artificial intelligence, which comprises the following steps: extracting features from text type demands, converting text information into numerical vector representation; performing semantic analysis on anesthetic demand data, extracting semantic information in the text, and displaying the extracted semantic information in the form of dynamic knowledge graph of entity-relation-entity structure; constructing the association relationship between demands; indexing each anesthetic demand data; constructing a subject ability value matrix based on the item response theory and outputting the anesthetic portrait; selecting collaborative filtering and content-based recommendation algorithm to generate a personalized recommendation list; and optimizing and improving the recommendation algorithm in real time. Through the dynamic knowledge graph, the automatic classification and efficient retrieval of anesthetic demands are realized, the classification accuracy and retrieval efficiency of the demands are improved, and the precise recommendation of anesthetists is realized through the construction of anesthetic portraits.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of artificial intelligence, in particular to an anesthesia scheme generation system based on big data analysis. BACKGROUND

[0002] With the rapid development of medical technology, the safety and effectiveness of surgical anesthesia are increasingly valued. Preoperative anesthesia visit is an important link to ensure the smooth progress of surgical anesthesia, which requires doctors to fully understand the patient's health status, medical history, medication, etc. in order to develop a personalized anesthesia plan. However, the traditional preoperative anesthesia visit method often relies on the experience and memory of doctors, and simple medical history inquiry and physical examination are performed on patients, which has the problems of incomplete information and missing important information, and cannot fully understand the patient's condition, including their physical function status, disease condition, medication, etc. to better predict their anesthesia risk.

[0003] In the prior art, anesthesia demand recommendations often rely on simple demand pushing or recommendations based on anesthesia history behavior, which often cannot accurately meet the individual needs of anesthetists. There is often a lack of effective correlation between anesthesia needs, making it difficult for anesthetists to form a systematic knowledge system during anesthesia. SUMMARY

[0004] To solve the above technical problems, the anesthesia scheme generation system based on big data analysis is provided, which solves the problems raised in the background technology.

[0005] To achieve the above purpose, the technical scheme adopted by the present application is:

[0006] The anesthesia scheme generation system based on big data analysis comprises:

[0007] A feature extraction module collects at least one existing anesthesia demand data and at least one anesthesia scheme, the anesthesia demand data and the anesthesia scheme are in a corresponding relationship, pre-processes them, and extracts features from text demands, converting text information into numerical vector representation;

[0008] A demand analysis module performs semantic analysis on anesthesia demand data, extracts semantic information in the text, and uses dependency syntax analysis to display the extracted semantic information in the form of a dynamic knowledge graph of entity-relation-entity structure, the semantic information including entities, relationships and attributes;

[0009] An association recognition module constructs the association relationship between demands based on the semantic analysis results, and uses the association relationship to realize the rapid positioning and mutual reference between demands;

[0010] An anesthesia optimization module indexes various anesthesia demand data and uses natural language processing technology to understand the anesthesia intent and optimize the anesthesia pathway.

[0011] The data processing module collects historical behavioral data of anesthesia protocols, records multi-dimensional data such as preference heatmaps and error distributions, constructs a subject competency value matrix based on item response theory, and outputs an anesthesia profile.

[0012] The recommendation generation module generates a personalized recommendation list based on anesthesia profiles and anesthesia demand data, using collaborative filtering and content-based recommendation algorithms.

[0013] The protocol production module optimizes and improves the recommendation algorithm in real time based on the effect of anesthesia feedback, and generates anesthesia protocols using a personalized recommendation list.

[0014] Preferably, the collection of historical behavioral data on anesthesia protocols, recording of preference heatmaps and error distribution data across multiple dimensions, and constructing a subject competency matrix based on item response theory and outputting it as an anesthesia profile specifically includes:

[0015] Data tracking technology is used to record the actual effects of anesthesia protocols during surgery.

[0016] Using heatmap analysis tools, the distribution of anesthesia preferences during surgery was recorded, and the focus of anesthesia attention was identified.

[0017] Collect error records of anesthesia protocols during practice or testing, including error types, error rates, and relevant knowledge points;

[0018] The preference heatmap is visualized to identify key areas and hot topics of interest in anesthesia, and anesthesia interest points and anesthesia preference characteristics are extracted based on the heatmap analysis results.

[0019] Statistical analysis of error records in anesthesia protocols during practice or testing is conducted to identify weaknesses and error-prone areas in the protocols.

[0020] Based on the error distribution analysis results, the difficulty and discrimination of each knowledge point are determined;

[0021] Using the Project Response Theory model, the capability value of the anesthesia protocol is calculated based on its performance during surgery.

[0022] The capability values ​​of the anesthesia protocol are organized according to knowledge points, a capability value matrix is ​​constructed, and the output is an anesthesia profile, which corresponds to the anesthesia protocol.

[0023] Preferably, the selecting a collaborative filtering and content-based recommendation algorithm to generate the personalized recommendation list based on the anesthesia image and anesthesia demand data specifically comprises:

[0024] The cosine similarity between the anesthesia schemes, the anesthesia demands and the anesthesia images is calculated respectively by using the cosine similarity formula;

[0025] Based on the historical data, a high similarity threshold is set;

[0026] It is judged whether the cosine similarity between the anesthesia schemes, the anesthesia demands or the anesthesia images is higher than the high similarity threshold, if yes, the anesthesia scheme group, the anesthesia demand group or the anesthesia image group is output as a similar anesthesia scheme, a similar anesthesia demand or a similar anesthesia image, if not, no output is made;

[0027] Based on the collaborative filtering algorithm, at least one recommendation list is generated from the perspective of the similar anesthesia schemes, the similar anesthesia demands and the similar anesthesia images, in which the similar anesthesia schemes are divided into the same category to obtain the anesthesia scheme category, the similar anesthesia demands are divided into the same category to obtain the anesthesia demand category, and the similar anesthesia images are divided into the same category to obtain the anesthesia image category;

[0028] According to the corresponding relationship between the anesthesia demand data and the anesthesia scheme, the anesthesia schemes of the same category are corresponded to the anesthesia demands of the same category and the anesthesia images of the same category;

[0029] Based on the historical behavior data of the anesthesia scheme and the popularity of the anesthesia demand, the personalized recommendation list is reordered.

[0030] Preferably, the generating the anesthesia scheme by using the personalized recommendation list specifically comprises:

[0031] The new anesthesia demand data is obtained, the cosine similarity between the new anesthesia demand data and the anesthesia demand in the recommendation list is calculated, the category of the anesthesia demand data with the maximum cosine similarity with the new anesthesia demand data in the recommendation list is selected as the anesthesia demand data target category;

[0032] The anesthesia scheme category corresponding to the anesthesia demand data target category is selected as the anesthesia scheme target category;

[0033] The anesthesia image with the maximum ability value in the anesthesia image category corresponding to the anesthesia scheme target category is selected as the target anesthesia image, and the anesthesia scheme corresponding to the target anesthesia image is selected as the scheme required by the new anesthesia demand data;

[0034] The calculation of the ability value of the anesthesia image is as follows:

[0035] When the ability value matrix of the anesthetic image is a square matrix, then the value of the determinant of the ability value matrix is calculated as the anesthetic image ability value;

[0036] When the ability value matrix of the anesthetic image is not a square matrix, then the ability value matrix of the anesthetic image is supplemented to a square matrix using 1 as an element and the original ability value matrix is replaced.

[0037] Compared with the prior art, the beneficial effects of the present application are that:

[0038] The automatic classification and efficient retrieval of anesthetic requirements are realized, the classification accuracy and retrieval efficiency of the requirements are greatly improved, the rapid positioning and mutual reference between the requirements are realized by using the association relationship, which helps the anesthetist to establish a systematic knowledge system during the anesthetic process and improve the anesthetic effect, the personalized recommendation list is generated by selecting the collaborative filtering and content-based recommendation algorithm based on the anesthetic image and anesthetic requirement data, the accurate recommendation for the anesthetist can be realized, and the utilization rate of the anesthetic requirement and the satisfaction of the anesthetist are improved. BRIEF DESCRIPTION OF DRAWINGS

[0039] Figure 1 The flowchart of the anesthetic scheme generation system based on big data analysis of the present application is shown in the figure;

[0040] Figure 2 The flowchart of the method for extracting features from text requirements and converting text information into numerical vector representation of the present application is shown in the figure;

[0041] Figure 3 The flowchart of the method for performing semantic analysis on anesthetic requirement data and extracting semantic information in the text of the present application is shown in the figure;

[0042] Figure 4 The flowchart of the method for constructing the association relationship between requirements of the present application is shown in the figure;

[0043] Figure 5 The flowchart of the method for establishing an index for each anesthetic requirement data of the present application is shown in the figure;

[0044] Figure 6 The flowchart of the method for constructing a subject ability value matrix and outputting it as an anesthetic image of the present application is shown in the figure;

[0045] Figure 7 The flowchart of the method for selecting collaborative filtering and content-based recommendation algorithm to generate a personalized recommendation list of the present application is shown in the figure;

[0046] Figure 8 The flowchart of the method for generating an anesthetic scheme using a personalized recommendation list of the present application is shown in the figure. DETAILED DESCRIPTION

[0047] The following description is used to disclose the present application to enable a person skilled in the art to implement the present application. The preferred embodiments in the following description are only as examples, and other obvious modifications can be conceived by those skilled in the art.

[0048] Referring to Figure 1 As shown in the figure, the anesthesia scheme generation system based on big data analysis includes:

[0049] A feature extraction module collects at least one existing anesthesia demand data and at least one anesthesia scheme, the anesthesia demand data and the anesthesia scheme are in a corresponding relationship, pre-processes them, and extracts features from text demands, and converts text information into a numerical vector representation;

[0050] A demand analysis module performs semantic analysis on anesthesia demand data, extracts semantic information in the text, uses dependency syntax analysis to display the extracted semantic information in the form of a dynamic knowledge graph of entity-relation-entity structure, and the semantic information includes entities, relationships and attributes;

[0051] An association recognition module constructs an association relationship between demands based on the semantic analysis results, and uses the association relationship to realize fast positioning and mutual reference between demands;

[0052] An anesthesia optimization module indexes each anesthesia demand data and uses natural language processing technology to understand anesthesia intent and optimize anesthesia path;

[0053] A data processing module collects historical behavior data of anesthesia schemes, records preference heat maps and error distribution multidimensional data, constructs a subject ability value matrix based on item response theory, and outputs an anesthesia portrait;

[0054] A recommendation generation module selects collaborative filtering and content-based recommendation algorithms to generate a personalized recommendation list based on the anesthesia portrait and anesthesia demand data;

[0055] A scheme production module optimizes and improves the recommendation algorithm in real time based on the effect of anesthesia feedback, and generates an anesthesia scheme using a personalized recommendation list.

[0056] Referring to Figure 2 Collect at least one existing anesthesia demand data and at least one anesthesia scheme, the anesthesia demand data and the anesthesia scheme are in a corresponding relationship, pre-process them, and extract features from text demands, and convert text information into a numerical vector representation, which specifically includes:

[0057] Remove irrelevant characters in anesthesia demand data text, the irrelevant characters include web tags, special symbols and spaces;

[0058] performing word segmentation on the Chinese text of the anesthesia demand data based on a dictionary-based word segmentation method to split the text into at least one independent word;

[0059] traversing the entire data set to collect all the appearing words and constructing a word list, each word in the word list being a feature candidate;

[0060] initializing a two-dimensional feature matrix, the rows representing text samples and the columns representing words in the word list, traversing each text sample to count the number of times each word appears in the sample and filling the number into the feature matrix;

[0061] based on the category distribution and the frequency of the word in the entire data set, calculating the expected frequency of each word in each category using an expected frequency formula;

[0062] for each word, calculating the chi-square value between the actual frequency and the expected frequency in each category using a chi-square value formula;

[0063] arranging the chi-square values of all words from large to small, determining the feature word quantity n based on the requirements of the specific task and the limitations of the calculation requirements, and selecting the first n words with the largest chi-square values as the feature words according to the sorting results;

[0064] constructing a text bag-of-words model, each word being represented as a one-hot vector with only one element being 1 and the rest being 0;

[0065] obtaining the numerical vector representation of the text information by adding the one-hot vectors of each word in the anesthesia demand data text.

[0066] The expected frequency formula is:

[0067]

[0068] wherein E ij is the expected frequency of word j in category i, N i is the number of texts in category i, T j is the number of times word j appears in the entire data set, and N is the number of texts in the entire data set.

[0069] The chi-square value formula is:

[0070]

[0071] wherein K j is the chi-square value of word j, O ij is the actual frequency of word j in category i, and k is the number of categories.

[0072] Referring to Figure 3As shown, the semantic analysis is performed on the anesthesia demand data, the semantic information in the text is extracted, the extracted semantic information is displayed in the form of dynamic knowledge graph of entity-relation-entity structure by using dependency syntax analysis, and the display specifically includes:

[0073] The preprocessed text is analyzed by using a dependency syntax analysis tool to reveal the dependency relationship between words, and the dependency relationship includes subject-predicate relationship, verb-object relationship and modifier-verb relationship.

[0074] The dependency syntax analysis result is output in a tree structure, each node represents a word, and each edge represents the dependency relationship between words.

[0075] In the dependency syntax analysis tree, noun phrases and nouns are identified as entity candidates, and the accuracy of the entities is confirmed based on part-of-speech tagging and context information.

[0076] The extracted entities are stored in an entity library.

[0077] In the dependency syntax analysis tree, verb phrases and verbs are identified as relation candidates, and the subject and object of the verb are analyzed to determine the semantic relationship between entities.

[0078] Based on context information and domain knowledge, the extracted relationships are verified and corrected.

[0079] Attribute extraction is achieved by analyzing the adjectives and modifiers in the dependency syntax tree.

[0080] An empty knowledge graph is created, the structure of the graph and the attributes of nodes and edges are defined, the attribute fields of entities and the attribute fields of relationships are set, the attribute fields of entities include name, type and description, and the attribute fields of relationships include relationship type and confidence.

[0081] The extracted entities and relationships are added to the knowledge graph, nodes are created for each entity, the attributes of the nodes are set, edges are created for each pair of entities with relationships, and the attributes of the edges are set.

[0082] Based on the addition of new text, new entities and relationships are extracted in real time, the knowledge graph is updated, new nodes and edges are added, and the attributes of existing nodes and edges are modified.

[0083] The knowledge graph is presented as a graphical interface, nodes are represented as circles, edges are represented as curves, and different colors, shapes and sizes are used to distinguish different types of entities and relationships.

[0084] Syntactic analysis is one of the key technologies in natural language processing, which is the process of analyzing the input text sentence to get the syntactic structure of the sentence. Analyzing the syntactic structure is not only the demand of language understanding itself, but also provides support for other natural language processing tasks. Dependency analysis is to identify the mutual dependency relationship between words in a sentence. Dependency syntactic analysis belongs to shallow syntactic analysis.

[0085] Referring to Figure 4 Based on the semantic analysis result, the association relationship between the requirements is constructed, and the rapid positioning and mutual reference between the requirements are realized by using the association relationship. Specifically, the method comprises the following steps:

[0086] Defining the core relationship between the requirements, the core relationship includes equivalence relationship, hierarchical relationship, association relationship, time sequence relationship, causal relationship, component part and attribute relationship;

[0087] Based on the semantic analysis result, the association relationship between the requirements is recognized;

[0088] The association relationship between the requirements is represented by using a graph data structure, the nodes in the graph represent the requirements, and the edges represent the association relationship. The weights of the edges are assigned based on the strength and type of the relationship;

[0089] Designing an anesthesia interface to search for requirements by entity, attribute or relationship. When the anesthesia requirement submits a query, the query condition is traversed and searched in the association relationship graph;

[0090] The query result is displayed in the form of a network graph, and the nodes and edges matching the query condition are highlighted;

[0091] When the anesthesia needs to reference the requirements, the system finds at least one other requirement related to the requirement according to the association relationship graph;

[0092] Tracking and managing the reference relationship between the requirements, when the content or location of the requirement changes, the reference relationship graph is updated in real time.

[0093] Designing a user-friendly query interface to allow anesthesia to search by entity name, attribute or relationship type. The query condition input by the anesthesia is parsed into a graph traversal query statement. When the anesthesia requirement submits a query, the system performs a depth-first search or a breadth-first search in the graph according to the parsed query condition, and filters out the nodes and edges that meet the condition according to the query condition.

[0094] Referring to Figure 5 Indexing each anesthesia requirement data, and using natural language processing technology to understand the anesthesia intent, and optimizing the anesthesia path. Specifically, the method comprises the following steps:

[0095] Initialize the index structure, design a data structure to store the inverted index, where the key is the word and the value is the list of each requirement that the word appears in;

[0096] Each requirement is represented in the list by a unique identifier of the requirement;

[0097] For each requirement, traverse its segmented text content;

[0098] Determine whether each word already exists in the index, if so, add the unique identifier of the requirement to the document list of the word, if not, create a new word entry and add the unique identifier of the requirement as the first element of the document list of the word entry;

[0099] Based on machine learning method, identify the query intent of anesthesia;

[0100] Based on the anesthesia intent and entity recognition results, rewrite or expand the query text;

[0101] Sort and filter the index in the anesthesia requirement database, calculate the similarity between the anesthesia requirement and the query text, and sort according to the similarity.

[0102] Use pre-trained machine anesthesia model to analyze the query text of anesthesia, identify the intent behind it, including information query (anesthesia wants to obtain specific information), recommendation request (anesthesia hopes to obtain the recommendation of related requirements) or problem solving (anesthesia is looking for requirements to solve specific problems).

[0103] Referring to Figure 6 As shown, collect historical behavior data of anesthesia plan, record preference heat map and error distribution multidimensional data, construct subject ability value matrix based on item response theory and output as anesthesia portrait, which specifically includes:

[0104] Record the actual effect of anesthesia plan during surgery through data burying technology;

[0105] Use heat map analysis tool to record the preference distribution of anesthesia plan during surgery, identify the focus content of anesthesia attention;

[0106] Collect error records of anesthesia plan in practice or test, including error type, error rate and involved knowledge points;

[0107] Visualize the preference heat map, identify the key areas and hot content of anesthesia attention, and extract anesthesia interest points and anesthesia preference features based on the heat map analysis results;

[0108] Statistical analysis of error records of anesthesia plan in practice or test, identify the weak points and error-prone points of anesthesia plan;

[0109] Determine the difficulty and discrimination of each knowledge point based on the error distribution analysis result;

[0110] According to the performance of the anesthesia scheme in the operation, the ability value of the anesthesia scheme is calculated by using the item response theory model;

[0111] The ability value of the anesthesia scheme is organized according to the knowledge points, the ability value matrix is constructed and output as the anesthesia portrait, and the anesthesia portrait corresponds to the anesthesia scheme.

[0112] It can capture and record various operation behaviors of anesthesia in operation, and ensure that anesthesia behaviors can be fully captured.

[0113] Referring to Figure 7 Based on the anesthesia portrait and the anesthesia demand data, the collaborative filtering and content-based recommendation algorithm are selected to generate the personalized recommendation list, which specifically includes:

[0114] The cosine similarity between the anesthesia schemes, the anesthesia demands and the anesthesia portraits is calculated by using the cosine similarity formula respectively;

[0115] Based on the historical data, a high similarity threshold is set;

[0116] Determine whether the cosine similarity between the anesthesia schemes or the anesthesia demands or the anesthesia portraits is higher than the high similarity threshold, if yes, output the group of anesthesia schemes or the group of anesthesia demands or the group of anesthesia portraits as similar anesthesia schemes or similar anesthesia demands or similar anesthesia portraits, if no, do not output;

[0117] Based on the collaborative filtering algorithm, at least one recommendation list is generated from the perspective of similar anesthesia schemes, similar anesthesia demands and similar anesthesia portraits, in which the similar anesthesia schemes are divided into the same category to obtain the anesthesia scheme category, the similar anesthesia demands are divided into the same category to obtain the anesthesia demand category, and the similar anesthesia portraits are divided into the same category to obtain the anesthesia portrait category;

[0118] According to the corresponding relationship between the anesthesia demand data and the anesthesia scheme, the anesthesia schemes of the same category are corresponded to the anesthesia demands of the same category and the anesthesia portraits of the same category;

[0119] Based on the historical behavior data of the anesthesia scheme and the popularity of the anesthesia demand, the personalized recommendation list is reordered.

[0120] The cosine similarity formula is:

[0121]

[0122] In the formula, C os is the cosine similarity between the anesthesia, the anesthesia demand and the anesthesia portrait and the anesthesia demand characteristics, respectively, are eigenvectors of the discipline capability value matrix of anesthesia or numerical vectors of the text information of the anesthesia demand data, respectively, are the lengths of the corresponding vectors.

[0123] Referring to Figure 8 As shown in the figure, the generation of the anesthesia scheme using the personalized recommendation list specifically includes:

[0124] Obtaining new anesthesia demand data, calculating the cosine similarity of the new anesthesia demand data and the anesthesia demand in the recommendation list, selecting the category of the anesthesia demand data with the largest cosine similarity with the new anesthesia demand data in the recommendation list as the target category of the anesthesia demand data;

[0125] The anesthesia scheme category corresponding to the target category of the anesthesia demand data is taken as the target category of the anesthesia scheme;

[0126] The anesthesia image with the largest capability value in the anesthesia image category corresponding to the target category of the anesthesia scheme is taken as the target anesthesia image, and the anesthesia scheme corresponding to the target anesthesia image is taken as the scheme required by the new anesthesia demand data;

[0127] The calculation of the capability value of the anesthesia image is as follows:

[0128] When the capability value matrix of the anesthesia image is a square matrix, the determinant value of the capability value matrix is calculated as the capability value of the anesthesia image;

[0129] When the capability value matrix of the anesthesia image is not a square matrix, 1 is used as an element to supplement the capability value matrix of the anesthesia image to a square matrix and replace the original capability value matrix.

[0130] When the scheme is screened, the scheme is screened according to the similarity, and the scheme similar to the demand is taken as the candidate. However, there are multiple candidate schemes, so further selection is required. When selecting, the scheme corresponding to the anesthesia image with the largest capability value is taken as the final scheme, and the effect of the scheme is the better scheme in the optional scheme.

[0131] Further, the present scheme further provides a computer readable storage medium having a computer readable program stored thereon, and the computer readable program is called to execute the above-mentioned anesthesia scheme generation system based on big data analysis.

[0132] It can be understood that the storage medium can be a magnetic medium, such as a floppy disk, a hard disk, a magnetic tape; an optical medium such as a DVD; or a semiconductor medium such as a solid state disk (SSD) and the like.

[0133] In summary, the application has the advantages that: automatic classification and efficient retrieval of anesthesia requirements are realized, the classification accuracy and retrieval efficiency of requirements are greatly improved, the quick positioning and mutual reference between requirements are realized by using the association relationship, which helps anesthetists to establish a systematic knowledge system in the anesthesia process, improves the anesthesia effect, the personalized recommendation list is generated by selecting the collaborative filtering and content-based recommendation algorithm based on the anesthesia portrait and anesthesia requirement data, the precise recommendation for anesthetists can be realized, and the utilization rate of anesthesia requirements and the satisfaction of anesthetists are improved.

[0134] The basic principles, main features and advantages of the present application are shown and described above. Those skilled in the art should understand that the present application is not limited to the above-mentioned embodiments, and the above-mentioned embodiments and descriptions in the specification are only the principles of the present application. Various changes and improvements can be made without departing from the spirit and scope of the present application, and these changes and improvements all fall within the scope of the claimed present application. The scope of protection required by the present application is defined by the appended claims and their equivalents.

Claims

1. Anesthesia protocol generation system based on big data analysis, characterized by, The system comprises: a feature extraction module that collects at least one existing anesthesia demand data and at least one anesthesia scheme, the anesthesia demand data and the anesthesia scheme being in a corresponding relationship, pre-processes the anesthesia demand data and the anesthesia scheme, and extracts features of the text type demand to convert the text information into a numerical vector representation; a demand analysis module that performs semantic analysis on the anesthesia demand data, extracts semantic information in the text, and uses dependency syntax analysis to present the extracted semantic information in the form of a dynamic knowledge graph of entity-relation-entity structure, the semantic information including entities, relations, and attributes; an association recognition module that constructs an association relationship between the demands based on the semantic analysis results and uses the association relationship to realize quick positioning and mutual reference between the demands; an anesthesia optimization module that indexes each anesthesia demand data and uses natural language processing technology to understand anesthesia intent and optimize anesthesia paths; a data processing module that collects historical behavior data of the anesthesia scheme, records preference heat maps and error distribution multidimensional data, constructs a subject ability value matrix based on the item response theory, and outputs the anesthesia portrait; a recommendation generation module that selects collaborative filtering and content-based recommendation algorithms to generate an individualized recommendation list based on the anesthesia portrait and the anesthesia demand data; a scheme production module that uses the individualized recommendation list to generate anesthesia schemes based on the effect of anesthesia feedback to optimize and improve the recommendation algorithm in real time; The collection of at least one existing anesthesia demand data and at least one anesthesia scheme, the anesthesia demand data and the anesthesia scheme being in a corresponding relationship, pre-processing, and feature extraction of the text type demand to convert the text information into a numerical vector representation specifically includes: removing irrelevant characters in the anesthesia demand data text, the irrelevant characters including webpage tags, special symbols, and spaces; performing word segmentation on the Chinese text of the anesthesia demand data based on a dictionary-based word segmentation method to split the text into at least one independent word; traversing the entire data set to collect all appearing words and constructing a word table, each word in the word table being a feature candidate; initializing a two-dimensional feature matrix, the rows representing text samples and the columns representing words in the word table, traversing each text sample to count the number of occurrences of each word in the sample, and filling the feature matrix; based on the category distribution and the frequency of the word in the entire data set, calculating the expected frequency of each word in each category using the expected frequency formula; for each word, calculating the chi-square value between the actual frequency and the expected frequency in each category using the chi-square value formula; sorting the chi-square values of all words from large to small, determining the feature word quantity n based on the requirements of the specific task and the limitations of the calculation requirements, and selecting the top n words with the largest chi-square values as the feature words according to the sorting results; constructing a text bag-of-words model, each word being represented as a one-hot vector with only one element being 1 and the rest being 0. The numerical vector representation of the text information is obtained by adding the one-hot vectors of each vocabulary in the anesthesia demand data text.

2. The big data analytics based anesthetic protocol generation system as claimed in claim 1, wherein, The semantic analysis of the anesthesia demand data extracts semantic information in the text, and uses dependency syntax analysis to present the extracted semantic information in the form of a dynamic knowledge graph of entity-relation-entity structure, which specifically includes: The preprocessed text is subjected to dependency syntax analysis using a dependency syntax analysis tool to reveal the dependency relationships between the vocabularies, including subject-predicate relationships, verb-object relationships, and modifier-noun relationships. The dependency syntax analysis results are output in a tree structure representation, with each node representing a vocabulary and each edge representing the dependency relationship between the vocabularies. In the dependency syntax analysis tree, noun phrases and nouns are identified as entity candidates, and the accuracy of the entities is confirmed based on part-of-speech tagging and contextual information. The extracted entities are stored in an entity library. In the dependency syntax analysis tree, verb phrases and verbs are identified as relation candidates, and the subject and object of the verb are analyzed to determine the semantic relationship between the entities. The extracted relationships are verified and corrected based on contextual information and domain knowledge. Attribute extraction is achieved by analyzing the adjectives and modifiers in the dependency syntax tree. An empty knowledge graph is created, and the structure of the graph and the attributes of the nodes and edges are defined. The entity attribute fields include name, type, and description, and the relationship attribute fields include relationship type and confidence. The extracted entities and relationships are added to the knowledge graph, and nodes are created for each entity with their attributes set. Edges are created for each pair of entities with a relationship, and the edge attributes are set. Based on the addition of new text, new entities and relationships are extracted in real time, the knowledge graph is updated, new nodes and edges are added, and the attributes of existing nodes and edges are modified. The knowledge graph is presented as a graphical interface, with nodes represented as circles and edges represented as curves. Different colors, shapes, and sizes are used to distinguish between different types of entities and relationships.

3. The big data analytics based anesthetic protocol generation system as claimed in claim 2, wherein, Based on the semantic analysis results, the association relationships between the demands are constructed, and the rapid positioning and mutual referencing between the demands are achieved using the association relationships, which specifically includes: Defining the core relationships between the demands, including equivalence, hierarchical, association, temporal, causal, component, and attribute relationships. Based on the semantic analysis results, the association relationships between the demands are identified. The association relationships between the demands are represented using a graph data structure, with nodes representing demands and edges representing association relationships. Based on the strength and type of the relationships, each edge is assigned a weight. Designing an anesthesia interface to search for demands through entities, attributes, or relationships. When an anesthesia demand is submitted for querying, the query conditions are traversed and searched in the association relationship graph. The query results are displayed in the form of a network graph, with the nodes and edges matching the query conditions highlighted. When an anesthesia demand needs to reference another demand, the system finds at least one other demand related to the demand based on the association relationship graph. The reference relationships between the demands are tracked and managed, and the reference relationship graph is updated in real time when the demand content or location changes.

4. The big data analytics based anesthetic protocol generation system of claim 3, wherein, The index of each anesthesia requirement data is established, and the anesthesia intention is understood by using natural language processing technology, and the anesthesia path is optimized, specifically including: Initializing the index structure, designing a data structure to store the inverted index, where the key is the word and the value is the list of each requirement where the word appears; The representation of each requirement in the list is the unique identifier of the requirement; For each requirement, traverse the text content after word segmentation; Determine whether each word already exists in the index, if so, add the unique identifier of the requirement to the document list of the word, if not, create a new word entry and add the unique identifier of the requirement as the first element of the document list of the word entry; Based on the method of machine learning, the query intention of anesthesia is recognized; Based on the anesthesia intention and entity recognition results, the query text is rewritten or expanded; Sort and filter the index in the anesthesia requirement database, calculate the similarity between the anesthesia requirement and the query text, and sort according to the similarity.

5. The big data analytics based anesthetic protocol generation system as claimed in claim 4, wherein, The historical behavior data of anesthesia scheme is collected, the preference heat map and error distribution multidimensional data are recorded, the discipline ability value matrix is constructed based on item response theory, and the anesthesia portrait is output, specifically including: Through data burying technology, record the actual effect of anesthesia scheme in surgery; Use heat map analysis tool to record the preference distribution of anesthesia scheme in surgery, identify the focus content of anesthesia attention; Collect error records of anesthesia scheme in practice or test, including error type, error rate and involved knowledge points; Visualize the preference heat map, identify the key areas and hot content of anesthesia attention, and extract anesthesia interest points and anesthesia preference features based on the heat map analysis results; Statistical analysis is made on the error records of anesthesia scheme in practice or test to identify the weak points and error-prone points of anesthesia scheme; Based on the error distribution analysis results, determine the difficulty and discrimination of each knowledge point; Use item response theory model to calculate the ability value of anesthesia scheme according to its performance in surgery; Organize the ability value of anesthesia scheme according to knowledge points, construct ability value matrix and output anesthesia portrait, which corresponds to anesthesia scheme.

6. The big data analytics based anesthetic protocol generation system as claimed in claim 5, wherein, Based on the anesthesia portrait and anesthesia requirement data, the collaborative filtering and content-based recommendation algorithm is used to generate personalized recommendation list, specifically including: Use cosine similarity formula to calculate the cosine similarity between anesthesia schemes, anesthesia requirements and anesthesia portraits respectively; Based on historical data, set a high similarity threshold; Determine whether the cosine similarity between anesthesia schemes or anesthesia requirements or anesthesia portraits is higher than the high similarity threshold, if so, output the group of anesthesia schemes or the group of anesthesia requirements or the group of anesthesia portraits as similar anesthesia schemes or similar anesthesia requirements or similar anesthesia portraits, if not, do not output. Based on the collaborative filtering algorithm, at least one recommendation list is generated from the perspective of similar anesthesia schemes, similar anesthesia needs and similar anesthesia images, in which similar anesthesia schemes are divided into the same category to obtain anesthesia scheme categories, similar anesthesia needs are divided into the same category to obtain anesthesia need categories, and similar anesthesia images are divided into the same category to obtain anesthesia image categories; According to the corresponding relationship between anesthesia need data and anesthesia schemes, the anesthesia schemes of the same category are corresponded to the anesthesia needs of the same category and the anesthesia images of the same category; Based on the historical behavior data of anesthesia schemes and the heat of anesthesia needs, the personalized recommendation list is reordered.

7. The big data analytics based anesthetic regimen generation system as claimed in claim 6, wherein, The use of the personalized recommendation list for anesthesia scheme generation specifically includes: Obtaining new anesthesia need data, calculating the cosine similarity of the new anesthesia need data and the anesthesia needs in the recommendation list, selecting the category of the anesthesia need data with the maximum cosine similarity to the new anesthesia need data in the recommendation list as the anesthesia need data target category; The anesthesia scheme category corresponding to the anesthesia need data target category is taken as the anesthesia scheme target category; The anesthesia image with the maximum ability value in the anesthesia image category corresponding to the anesthesia scheme target category is taken as the target anesthesia image, and the anesthesia scheme corresponding to the target anesthesia image is taken as the scheme required by the new anesthesia need data; The calculation of the ability value of the anesthesia image is as follows: When the ability value matrix of the anesthesia image is a square matrix, the value of the determinant of the ability value matrix is taken as the ability value of the anesthesia image; When the ability value matrix of the anesthesia image is not a square matrix, 1 is used as an element, the ability value matrix of the anesthesia image is supplemented to a square matrix and the original ability value matrix is replaced.

Images