Traditional Chinese Medicine Health Management Cloud Platform Based on Big Data

Abstract

The present invention relates to the technical field of medical data anomaly detection, and specifically relates to a traditional Chinese medicine health management cloud platform based on big data, including a processor and a memory. The processor is used to process instructions stored in the memory to implement the following steps: obtaining the initial diagnosis and treatment information corresponding to the patient to be detected and the self-reported information to be detected, as well as the initial diagnosis and treatment information corresponding to historical patients and all historical self-reported information; screening out similar patients; extracting keywords from the self-reported information and clustering the keywords; constructing a data change sequence corresponding to the target cluster and a target representative data; performing anomaly detection on the target representative data according to the similarity between the target representative data and the elements in its corresponding data change sequence, so as to realize the anomaly detection of the self-reported information to be detected. The present invention realizes the anomaly detection of the self-reported information by processing the diagnosis and treatment information and the self-reported information, and improves the accuracy of anomaly detection of the patient's self-reported information.

Description

Technical Field

[0001] The present invention relates to the technical field of medical data anomaly detection, and particularly to a traditional Chinese medicine health management cloud platform based on big data. Background Art

[0002] Traditional Chinese medicine health management is a comprehensive service that combines traditional Chinese medicine theory with modern health management methods. Therefore, during the process of traditional Chinese medicine health management, it is often necessary to manage the collected patient diagnosis and treatment information and patient self-reported information. Among them, the patient diagnosis and treatment information can be the test result information obtained by the patient through testing equipment before seeing a traditional Chinese medicine doctor. For example, the patient diagnosis and treatment information can include: test results such as blood routine, urine routine, and gastric ultrasound. The patient self-reported information can represent the patient's description of their own disease symptoms. Due to the differences in patients' self-awareness, some patients may have difficulty accurately describing their own disease symptoms, and incorrect descriptions of disease symptoms often cause trouble to subsequent disease analysis. Therefore, during the process of traditional Chinese medicine health management, it is often necessary to perform anomaly detection on the collected patient self-reported information.

[0003] Currently, when performing anomaly detection on data, the commonly used method is to use regular expressions to perform anomaly detection on the data. However, when using regular expressions to perform anomaly detection on patient self-reported information, the following technical problems often exist:

[0004] Through regular expressions, format error anomalies in patient self-reported information can often be detected, such as inconsistent medical record numbers and date formats, etc., but it is often difficult to accurately detect incorrect descriptions of patient disease symptoms or typos in records, resulting in poor accuracy of anomaly detection for patient self-reported information. Summary of the Invention

[0005] In order to solve the technical problem of poor accuracy of anomaly detection for patient self-reported information, the present invention proposes a traditional Chinese medicine health management cloud platform based on big data.

[0006] In a first aspect, the present invention provides a traditional Chinese medicine health management cloud platform based on big data, including a processor and a memory. The processor is used to process the instructions stored in the memory to implement the following steps:

[0007] Obtain the initial diagnosis and treatment information and the to-be-detected self-reported information corresponding to the patient to be detected, as well as the initial diagnosis and treatment information corresponding to each historical patient and all historical self-reported information;

[0008] According to the similarity between the initial diagnosis and treatment information corresponding to the patient to be detected and that of all historical patients, screen out similar patients from all historical patients;

[0009] Extract keywords from the self-reported information to be detected and all historical self-reported information, and cluster all the extracted keywords to obtain target clusters;

[0010] Construct a data change sequence corresponding to the target cluster according to the keywords belonging to the same target cluster in all historical self-reported information of all similar patients;

[0011] Construct target representative data corresponding to the target cluster according to the keywords belonging to the same target cluster in the self-reported information to be detected of the patient to be detected;

[0012] Perform anomaly detection on each target representative data according to the similarity between each target representative data and the elements in the data change sequence corresponding to its target cluster, so as to realize the anomaly detection of the self-reported information to be detected.

[0013] Combined with the first aspect above, in a possible implementation manner, the screening of similar patients from all historical patients according to the similarity between the initial diagnosis and treatment information of the patient to be detected and that of all historical patients includes:

[0014] Based on each initial diagnosis and treatment information, construct a target representative matrix corresponding to each initial diagnosis and treatment information;

[0015] Determine the cosine similarity between the target representative matrix of the patient to be detected and the target representative matrix of each historical patient as the target similarity between the patient to be detected and each historical patient;

[0016] Screen out historical patients with a target similarity greater than a preset similarity threshold with the patient to be detected from all historical patients as similar patients.

[0017] Combined with the first aspect above, in a possible implementation manner, the clustering of all the extracted keywords to obtain target clusters includes:

[0018] Construct a target knowledge graph with the extracted keywords as entities;

[0019] Determine the semantic relevance between every two keywords according to the distance between every two keywords in the target knowledge graph, where the distance between different keywords in the target knowledge graph is negatively correlated with their semantic relevance;

[0020] Cluster all keywords according to the semantic relevance between different keywords, and determine the clustered clusters as target clusters.

[0021] Combined with the above first aspect, in a possible implementation manner, constructing a data change sequence corresponding to the target cluster according to the keywords belonging to the same target cluster in all historical self-report information corresponding to all similar patients includes:

[0022] Screen out the historical self-report information with the same corresponding treatment stage from all historical self-report information corresponding to all similar patients to form a group of historical self-report information under each treatment stage;

[0023] Determine any one target cluster as the marked cluster, and construct reference representative data for each treatment stage under the marked cluster according to all the keywords belonging to the marked cluster in the group of historical self-report information under each treatment stage;

[0024] Construct the data change sequence corresponding to the marked cluster from the reference representative data for all treatment stages under the marked cluster.

[0025] Combined with the above first aspect, in a possible implementation manner, constructing reference representative data for each treatment stage under the marked cluster according to all the keywords belonging to the marked cluster in the group of historical self-report information under each treatment stage includes:

[0026] Obtain the word vectors corresponding to each keyword, and determine the mean value of the word vectors corresponding to all the keywords belonging to the marked cluster in the group of historical self-report information under each treatment stage as the reference representative data for each treatment stage under the marked cluster.

[0027] Combined with the above first aspect, in a possible implementation manner, constructing target representative data corresponding to the target cluster according to the keywords belonging to the same target cluster in the to-be-detected self-report information corresponding to the to-be-detected patient includes:

[0028] Determine any one target cluster as the marked cluster, and obtain the word vectors corresponding to each keyword;

[0029] Determine the mean value of the word vectors corresponding to all the keywords belonging to the marked cluster in the to-be-detected self-report information as the target representative data corresponding to the marked cluster.

[0030] Combined with the above first aspect, in a possible implementation manner, performing anomaly detection on each target representative data according to the similarity between each target representative data and the elements in the data change sequence corresponding to its corresponding target cluster includes:

[0031] Determine the target treatment stage corresponding to each target representative data according to the cosine similarity between each target representative data and the reference representative data in the data change sequence corresponding to its corresponding target cluster;

[0032] Determine the mean of the start times of the target treatment phases corresponding to all target representative data as the treatment representative time;

[0033] Perform anomaly detection on each target representative data according to the difference between the treatment representative time and the start time of the target treatment phase corresponding to each target representative data.

[0034] Combined with the above first aspect, in a possible implementation manner, the determining the target treatment phase corresponding to each target representative data according to the cosine similarity between each target representative data and the reference representative data in the data change sequence corresponding to its corresponding target cluster includes:

[0035] Determine any one target representative data as the marked representative data, determine the target cluster corresponding to the marked representative data as the temporary cluster, and determine each reference representative data in the data change sequence corresponding to the temporary cluster as the temporary representative data;

[0036] Screen out the temporary representative data with the largest cosine similarity with the marked representative data from all temporary representative data as the comparison representative data;

[0037] Set the treatment phase corresponding to the comparison representative data as the target treatment phase corresponding to the marked representative data.

[0038] Combined with the above first aspect, in a possible implementation manner, the performing anomaly detection on each target representative data according to the difference between the treatment representative time and the start time of the target treatment phase corresponding to each target representative data includes:

[0039] Normalize the absolute value of the difference between the treatment representative time and the start time of the target treatment phase corresponding to each target representative data to obtain the abnormal deviation degree corresponding to each target representative data;

[0040] If the abnormal deviation degree corresponding to the target representative data is greater than the preset abnormal threshold, determine that the target representative data is abnormal.

[0041] Combined with the above first aspect, in a possible implementation manner, the anomaly detection method for the self-reported information to be detected includes:

[0042] If there is an abnormal target representative data, determine that the self-reported information to be detected is abnormal.

[0043] The present invention has the following beneficial effects:

[0044] The cloud platform for traditional Chinese medicine health management based on big data of the present invention processes diagnosis and treatment information and self-reported information, realizes the anomaly detection of self-reported information, solves the technical problem of poor accuracy in anomaly detection of patients' self-reported information, and improves the accuracy of anomaly detection of patients' self-reported information. Compared with anomaly detection through regular expressions, when the present invention performs anomaly detection on the self-reported information to be detected corresponding to the patient to be detected, the disease conditions of patients with similar conditions often have a certain correlation. Therefore, similar patients with similar diagnosis and treatment conditions to the patient to be detected are selected from all historical patients, and the data change sequence representing the overall situation of all similar patients and the target representative data representing the overall situation of the patient to be detected are quantified under the same target cluster. By analyzing the similarity between each target representative data and the elements in the corresponding data change sequence of the target cluster, the anomaly detection of the self-reported information to be detected is realized, thereby improving the accuracy of anomaly detection of the self-reported information to be detected. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0046] Figure 1 It is a flowchart of the method implemented by the cloud platform for traditional Chinese medicine health management based on big data of the present invention;

[0047] Figure 2 It is a flowchart of the method for obtaining the target treatment stage of the present invention;

[0048] Figure 3 It is a flowchart of the anomaly detection of the target representative data of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0049] In order to further elaborate on the technical means and effects adopted by the present invention to achieve the intended invention purpose, the following will, in conjunction with the accompanying drawings and preferred embodiments, describe in detail the specific implementation manners, structures, features, and effects of the technical solutions proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0050] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the technical field to which the present invention belongs.

[0051] Refer toFigure 1 shows a method flowchart implemented by a big data-based traditional Chinese medicine health management cloud platform according to the present invention. Specifically, the big data-based traditional Chinese medicine health management cloud platform includes a processor and a memory, and the processor is configured to process instructions stored in the memory to implement the following steps:

[0052] Step S1, obtain the initial diagnosis and treatment information corresponding to the patient to be detected and the self-reported information to be detected, as well as the initial diagnosis and treatment information corresponding to each historical patient and all historical self-reported information.

[0053] Among them, the patient to be detected can be a patient to be detected for abnormal self-reported information. The initial diagnosis and treatment information can characterize the condition of the patient during the first visit for traditional Chinese medicine treatment. The initial diagnosis and treatment information can include: the examination results that assist traditional Chinese medicine in judgment during the first visit of the patient for traditional Chinese medicine treatment, and the diagnosis results of traditional Chinese medicine during the first visit. For example, the initial diagnosis and treatment information can include, but is not limited to: blood routine examination results, urine routine examination results, gastric ultrasound examination results, and traditional Chinese medicine diagnosis results. It should be noted that during traditional Chinese medicine treatment, traditional Chinese medicine often observes the examination results obtained by the patient through corresponding equipment to assist in treatment judgment. The examination results obtained through equipment examination can be, but are not limited to: blood routine examination results, urine routine examination results, and gastric ultrasound examination results. The self-reported information to be detected can be the information obtained by the patient to be detected describing his own disease symptoms. The historical patient can be a patient who has received traditional Chinese medicine treatment in the past. The historical self-reported information can be the information obtained by the historical patient describing his own disease symptoms.

[0054] As an example, it is possible to collect the blood routine, urine routine, gastric ultrasound and other examination results that assist traditional Chinese medicine diagnosis during the first visit of the patient to be detected during traditional Chinese medicine treatment, and record the diagnosis results of traditional Chinese medicine for the patient to be detected during the first visit of the patient to be detected during traditional Chinese medicine treatment. These examination results and traditional Chinese medicine diagnosis results constitute the initial diagnosis and treatment information corresponding to the patient to be detected. It is possible to collect the information describing the patient's own disease symptoms during the most recent visit of the patient to be detected during traditional Chinese medicine treatment to constitute the self-reported information to be detected corresponding to the patient to be detected. Similarly, it is possible to collect the initial diagnosis and treatment information corresponding to each historical patient and all historical self-reported information.

[0055] Step S2, screen out similar patients from all historical patients according to the similarity between the initial diagnosis and treatment information corresponding to the patient to be detected and all historical patients.

[0056] It should be noted that the more similar patients are screened out, the relatively better the effect of subsequent abnormal detection of self-reported information will be.

[0057] As an example, this step may include the following steps:

[0058] The first step, constructing a target representative matrix corresponding to each initial diagnosis and treatment information may include the following sub-steps:

[0059] The first sub-step, for the examination results with numerical data in the initial diagnosis and treatment information, the vector composed of all numerical data representing different indicators in each examination result can be denoted as the initial examination vector.

[0060] For example, the vector composed of all numerical data representing different indicators in the urine routine examination results can be used as an initial examination vector.

[0061] The second sub-step, for the text information in the initial diagnosis and treatment information such as the traditional Chinese medicine diagnosis result, a keyword library can be manually constructed according to expert suggestions to obtain different one-hot vectors corresponding to different keywords; then, the words in the traditional Chinese medicine diagnosis result are matched with the keywords in the constructed keyword library to obtain a matching keyword sequence, and the one-hot vectors of all keywords in the matching keyword sequence are superimposed to obtain the feature vector of the traditional Chinese medicine diagnosis result.

[0062] The third sub-step, the matrix composed of all initial examination vectors corresponding to each initial diagnosis and treatment information and the feature vector of the traditional Chinese medicine diagnosis result is used as the target representative matrix corresponding to each initial diagnosis and treatment information.

[0063] For example, each initial examination vector can be used as a row of the matrix, and the feature vector of the traditional Chinese medicine diagnosis result can be used as a row of the matrix. Based on the longest row vector in the matrix, the other row vectors are filled with 0s to make the lengths of all row vectors in the matrix the same, and the finally obtained matrix is denoted as the target representative matrix. Each row in the matrix can be called a row vector.

[0064] It should be noted that the target representative matrix corresponding to the initial diagnosis and treatment information can represent the patient's condition symptoms corresponding to the initial diagnosis and treatment information.

[0065] The second step, the cosine similarity between the target representative matrix corresponding to the to-be-detected patient and the target representative matrix corresponding to each historical patient is determined as the target similarity between the to-be-detected patient and each historical patient.

[0066] Among them, the target representative matrix corresponding to the to-be-detected patient is the target representative matrix corresponding to the initial diagnosis and treatment information corresponding to the to-be-detected patient. The target representative matrix corresponding to the historical patient is the target representative matrix corresponding to the initial diagnosis and treatment information corresponding to the historical patient.

[0067] The third step, select historical patients from all historical patients whose target similarity with the to-be-detected patient is greater than the preset similarity threshold as similar patients.

[0068] Among them, the preset similarity threshold can be the maximum similarity that is preset and considered when two patients are not similar. For example, the preset similarity threshold can be 0.7.

[0069] It should be noted that the disease symptoms of similar patients at the initial stage of traditional Chinese medicine treatment are often similar to those of the patient to be detected at the initial stage of traditional Chinese medicine treatment.

[0070] Step S3: Extract keywords from the self-reported information to be detected and all historical self-reported information, and cluster all the extracted keywords to obtain the target clusters.

[0071] As an example, this step may include the following steps:

[0072] The first step is to extract keywords from the self-reported information to be detected and all historical self-reported information.

[0073] For example, a keyword library can be set manually, and the words in the self-reported information to be detected and all historical self-reported information are matched one by one with the keywords in the manually set keyword library. The words that can be matched in the self-reported information to be detected and all historical self-reported information are used as the extracted keywords.

[0074] The second step is to construct a target knowledge graph with the extracted keywords as entities.

[0075] For example, the keywords extracted from the self-reported information to be detected and all historical self-reported information can be used as entities to construct a knowledge graph, and the constructed knowledge graph is used as the target knowledge graph.

[0076] It should be noted that the relevant descriptions of the disease in the self-reported information to be detected and all historical self-reported information can be collected, and the entities of the descriptions are identified. Specifically, the extracted keywords can be used as entities, and relationship extraction is performed to build a knowledge graph network. At this time, the built knowledge graph is the target knowledge graph.

[0077] The third step is to determine the semantic relevance between every two keywords according to the distance between every two keywords in the above-mentioned target knowledge graph.

[0078] Among them, the distance between two keywords in the target knowledge graph is also the shortest path length between these two keywords in the target knowledge graph. The distance between different keywords in the target knowledge graph can be negatively correlated with their semantic relevance.

[0079] It should be noted that in actual situations, the correlation between one-hot vectors corresponding to different keywords is often 0, but in fact, there may be semantic connections between them. For example, if two keywords are stomachache and acid reflux in sequence, the correlation between the one-hot vectors corresponding to stomachache and acid reflux is often 0, but in fact, both stomachache and acid reflux describe the situation of the stomach, there is a semantic connection, and their distance in the knowledge graph is relatively smaller. Therefore, when the distance between two keywords in the target knowledge graph is smaller, it often indicates that the semantic correlation between these two keywords is greater.

[0080] For example, the formula for determining the semantic correlation between two keywords can be:

[0081] ; is the i th keyword and the j th keyword's semantic correlation. i and j are the serial numbers of different keywords. is the i th keyword and the j th keyword's distance in the target knowledge graph.

[0082] Fourth step, according to the semantic correlation between different keywords, cluster all keywords, and determine the clustering clusters obtained by clustering as the target clusters.

[0083] For example, all keywords can be hierarchically clustered according to the semantic correlation between different keywords, and the clustering clusters obtained by clustering are denoted as target clusters.

[0084] It should be noted that when the semantic correlation between two keywords is greater, these two keywords are often classified into the same clustering cluster.

[0085] Step S4, according to the keywords belonging to the same target cluster in all historical self-report information corresponding to all similar patients, construct the data change sequence corresponding to the target cluster.

[0086] As an example, this step may include the following steps:

[0087] First step, screen out the historical self-report information with the same corresponding treatment stage from all historical self-report information corresponding to all similar patients to form a group of historical self-report information for each treatment stage.

[0088] Among them, the treatment stage can be a pre-set treatment time period, and its corresponding duration can be a pre-set duration. For example, if the duration corresponding to the treatment stage is 1 day, the first day of the patient's traditional Chinese medicine treatment can be a treatment stage, the second day of the patient's traditional Chinese medicine treatment can be a treatment stage, the third day of the patient's traditional Chinese medicine treatment can be a treatment stage, and so on. All treatment stages corresponding to all similar patients can be obtained. The method for obtaining the historical self-report information group under the treatment stage can be: taking any treatment stage as the marked treatment stage, and constructing the historical self-report information group under the marked treatment stage from the historical self-report information collected within the marked treatment stage among all the historical self-report information corresponding to all similar patients.

[0089] It should be noted that the more treatment stages there are corresponding to all similar patients, the relatively more accurate the results obtained by subsequent self-report information anomaly detection based on all the treatment stages corresponding to all similar patients will be.

[0090] In the second step, take any target cluster as the marked cluster, and construct the reference representative data of each treatment stage under the marked cluster according to all the keywords belonging to the above-mentioned marked cluster in the historical self-report information group under each treatment stage.

[0091] For example, through Word2Vec, the word vector corresponding to each keyword can be obtained, and the mean value of the word vectors corresponding to all the keywords belonging to the above-mentioned marked cluster in the historical self-report information group under each treatment stage is determined as the reference representative data of each treatment stage under the above-mentioned marked cluster. Among them, Word2Vec is a word embedding model based on neural networks, which can map vocabulary into real number vectors, denoted as word vectors.

[0092] In the third step, the reference representative data of all treatment stages under the above-mentioned marked cluster are formed into the data change sequence corresponding to the above-mentioned marked cluster.

[0093] Step S5, construct the target representative data corresponding to the target cluster according to the keywords belonging to the same target cluster in the to-be-detected self-report information corresponding to the to-be-detected patient.

[0094] As an example, this step may include the following steps:

[0095] In the first step, take any target cluster as the marked cluster, and through Word2Vec, obtain the word vector corresponding to each keyword.

[0096] In the second step, determine the mean value of the word vectors corresponding to all the keywords belonging to the above-mentioned marked cluster in the to-be-detected self-report information as the target representative data corresponding to the above-mentioned marked cluster.

[0097] Step S6: Based on the similarity between each target representative data and the elements in the corresponding data change sequence of its target cluster, perform anomaly detection on each target representative data, thereby realizing the anomaly detection of the self-reported information to be detected.

[0098] As an example, this step may include the following steps:

[0099] First step: Based on the cosine similarity between each target representative data and the reference representative data in the corresponding data change sequence of its target cluster, determine the target treatment stage corresponding to each target representative data.

[0100] As Figure 2 shown, determining the target treatment stage corresponding to each target representative data may include the following steps:

[0101] Step 201: Designate any one target representative data as the marked representative data, and designate the target cluster corresponding to the marked representative data as the temporary cluster. Designate each reference representative data in the data change sequence corresponding to the temporary cluster as the temporary representative data.

[0102] Step 202: Screen out the temporary representative data with the largest cosine similarity to the marked representative data from all the temporary representative data as the comparison representative data.

[0103] Step 203: Set the treatment stage corresponding to the comparison representative data as the target treatment stage corresponding to the marked representative data.

[0104] Second step: Determine the mean of the start times of the target treatment stages corresponding to all the target representative data as the treatment representative time.

[0105] Third step: Based on the difference between the treatment representative time and the start time of the target treatment stage corresponding to each target representative data, perform anomaly detection on each target representative data.

[0106] As Figure 3 shown, performing anomaly detection on each target representative data may include the following steps:

[0107] Step 301: Normalize the absolute value of the difference between the treatment representative time and the start time of the target treatment stage corresponding to each target representative data to obtain the anomaly deviation degree corresponding to each target representative data.

[0108] It should be noted that patients' descriptions of most of their own disease symptoms are often relatively accurate. That is to say, the accurate symptom description data in the self-reported information often exceeds the abnormal description data. Therefore, the treatment representative moment can often represent the start moment of the current treatment stage of the patient to be detected. At this time, if the start moment of the target treatment stage corresponding to the target representative data deviates more from the treatment representative moment, it often indicates that the disease symptom description represented by the target representative data is more likely to be abnormal and requires reminding the patient or the traditional Chinese medicine doctor to conduct key verification.

[0109] Step 302, if the abnormal deviation degree corresponding to the target representative data is greater than the preset abnormal threshold, it is determined that the target representative data is abnormal.

[0110] Among them, the preset abnormal threshold can be a threshold set in advance, and it can be 0.6.

[0111] It should be noted that if there is no reference representative data in the data change sequence corresponding to the temporary cluster, the abnormal deviation degree corresponding to the target representative data can be set to 1.

[0112] Fourth step, if there is abnormal target representative data, it is determined that the self-reported information to be detected is abnormal.

[0113] It should be noted that when the self-reported information to be detected is abnormal, it is often necessary to remind the patient or the traditional Chinese medicine doctor to conduct key verification on the abnormal target representative data in the self-reported information to be detected.

[0114] In summary, compared with anomaly detection through regular expressions, when the present invention performs anomaly detection on the self-reported information to be detected corresponding to the patient to be detected, the disease conditions of patients with similar diseases often have a certain correlation. Therefore, similar patients with similar diagnosis and treatment conditions to the patient to be detected are selected from all historical patients, and the data change sequence representing the overall conditions of all similar patients and the target representative data representing the overall conditions of the patient to be detected in the same target cluster are quantified. By analyzing the similarity between each target representative data and the elements in the data change sequence corresponding to its target cluster, the anomaly detection of the self-reported information to be detected is realized, thereby improving the accuracy of anomaly detection of the self-reported information to be detected.

[0115] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the various embodiments of the present invention, and should all be included in the protection scope of the present invention.

Claims

1. A traditional Chinese medicine health management cloud platform based on big data, characterized in that, It includes a processor and a memory. The processor is used to process the instructions stored in the memory to implement the following steps: Obtain the initial diagnosis and treatment information corresponding to the patient to be detected and the self-reported information to be detected, as well as the initial diagnosis and treatment information corresponding to each historical patient and all historical self-reported information; According to the similarity between the initial diagnosis and treatment information corresponding to the patient to be detected and that of all historical patients, screen out similar patients from all historical patients; Extract keywords from the self-reported information to be detected and all historical self-reported information, and cluster all the extracted keywords to obtain target clusters; According to the keywords belonging to the same target cluster in all historical self-reported information corresponding to all similar patients, construct a data change sequence corresponding to the target cluster; According to the keywords belonging to the same target cluster in the self-reported information to be detected corresponding to the patient to be detected, construct a target representative data corresponding to the target cluster; According to the similarity between each target representative data and the elements in the data change sequence corresponding to its corresponding target cluster, perform anomaly detection on each target representative data, so as to realize the anomaly detection of the self-reported information to be detected; The constructing a data change sequence corresponding to the target cluster according to the keywords belonging to the same target cluster in all historical self-reported information corresponding to all similar patients includes: Screen out the historical self-reported information with the same corresponding treatment stage from all historical self-reported information corresponding to all similar patients, and form a group of historical self-reported information under each treatment stage; Determine any one target cluster as a marked cluster, and construct reference representative data for each treatment stage under the marked cluster according to all the keywords belonging to the marked cluster in the group of historical self-reported information under each treatment stage; Form the data change sequence corresponding to the marked cluster with the reference representative data for all treatment stages under the marked cluster; The performing anomaly detection on each target representative data according to the similarity between each target representative data and the elements in the data change sequence corresponding to its corresponding target cluster includes: Determine the target treatment stage corresponding to each target representative data according to the cosine similarity between each target representative data and the reference representative data in the data change sequence corresponding to its corresponding target cluster; Determine the mean value of the start times of the target treatment stages corresponding to all target representative data as the treatment representative time; Perform anomaly detection on each target representative data according to the difference between the treatment representative time and the start time of the target treatment stage corresponding to each target representative data.

2. The traditional Chinese medicine health management cloud platform based on big data according to claim 1, wherein, The screening out similar patients from all historical patients according to the similarity between the initial diagnosis and treatment information corresponding to the patient to be detected and that of all historical patients includes: Based on each initial diagnosis and treatment information, construct a target representative matrix corresponding to each initial diagnosis and treatment information; Determine the cosine similarity between the target representative matrix corresponding to the patient to be detected and the target representative matrix corresponding to each historical patient as the target similarity between the patient to be detected and each historical patient; Screen out historical patients from all historical patients whose target similarity with the patient to be detected is greater than a preset similarity threshold as similar patients.

3. The cloud platform for traditional Chinese medicine health management based on big data according to claim 1, characterized in that Clustering all the extracted keywords to obtain target clusters, including: Constructing a target knowledge graph with the extracted keywords as entities; Determining the semantic relevance between every two keywords according to the distance between them in the target knowledge graph, where the distance between different keywords in the target knowledge graph is negatively correlated with their semantic relevance; Clustering all the keywords according to the semantic relevance between different keywords, and determining the clustering clusters obtained by clustering as target clusters.

4. A traditional Chinese medicine health management cloud platform based on big data according to claim 1, characterized in that, Constructing reference representative data for each treatment stage under the marked cluster according to all the keywords in the historical self-report information group belonging to the marked cluster in each treatment stage, including: Obtaining the word vector corresponding to each keyword, and determining the mean value of the word vectors corresponding to all the keywords in the historical self-report information group belonging to the marked cluster in each treatment stage as the reference representative data for each treatment stage under the marked cluster.

5. The traditional Chinese medicine health management cloud platform based on big data according to claim 1, wherein, Constructing target representative data corresponding to the target cluster according to the keywords belonging to the same target cluster in the to-be-detected self-report information of the to-be-detected patient, including: Determining any one target cluster as the marked cluster, and obtaining the word vector corresponding to each keyword; Determining the mean value of the word vectors corresponding to all the keywords in the to-be-detected self-report information belonging to the marked cluster as the target representative data corresponding to the marked cluster.

6. The traditional Chinese medicine health management cloud platform based on big data according to claim 1, characterized in that Determining the target treatment stage corresponding to each target representative data according to the cosine similarity between each target representative data and the reference representative data in the data change sequence corresponding to its corresponding target cluster, including: Determining any one target representative data as the marked representative data, determining the target cluster corresponding to the marked representative data as the temporary cluster, and determining each reference representative data in the data change sequence corresponding to the temporary cluster as the temporary representative data; Screening out the temporary representative data with the largest cosine similarity with the marked representative data from all the temporary representative data as the comparison representative data; Setting the treatment stage corresponding to the comparison representative data as the target treatment stage corresponding to the marked representative data.

7. A traditional Chinese medicine health management cloud platform based on big data according to claim 1, characterized in that, Performing anomaly detection on each target representative data according to the difference between the treatment representative moment and the start moment of the target treatment stage corresponding to each target representative data, including: Normalizing the absolute value of the difference between the treatment representative moment and the start moment of the target treatment stage corresponding to each target representative data to obtain the anomaly deviation degree corresponding to each target representative data; If the anomaly deviation degree corresponding to the target representative data is greater than the preset anomaly threshold, determining that the target representative data is abnormal.

8. The cloud platform for traditional Chinese medicine health management based on big data according to claim 7, characterized in that, Anomaly detection method for to-be-detected self-report information, including: If there is abnormal target representative data, determining that the to-be-detected self-report information is abnormal.

Images