Search

CN-121983296-A - Intelligent medical assistance method and system based on big data analysis

CN121983296ACN 121983296 ACN121983296 ACN 121983296ACN-121983296-A

Abstract

The invention discloses an intelligent medical auxiliary method and system based on big data analysis, which relate to the technical field of medical treatment and aim at the problems that multi-source medical data are difficult to uniformly analyze and cross-source characteristic association cannot be dynamically mined to predict disease development risk, and constructing a disease development prediction path based on the association relation set, obtaining a trend sequence, screening high-risk nodes from the trend sequence, merging multiple factors to calculate potential risk values to generate a risk assessment list, finally generating a comprehensive report according to priority ranking, outputting an auxiliary judgment result, and dynamically optimizing the result by merging new data, thereby realizing the technical effects of automatically discovering deep association from multi-source heterogeneous data and dynamically predicting the disease risk.

Inventors

  • YU LINGZHI
  • CHEN YANG

Assignees

  • 湖南湘约健康科技有限公司

Dates

Publication Date
20260505
Application Date
20260408

Claims (10)

  1. 1. An intelligent medical assistance method based on big data analysis is characterized by comprising the following steps: s100, acquiring multi-source medical data, and converting the multi-source medical data into a feature representation with uniform dimension through a data processing module to obtain a standardized medical data set; S200, grouping the features in the standardized medical data set by adopting a clustering algorithm to obtain a plurality of feature clusters; S300, extracting co-occurrence feature modes of cross-data sources from a plurality of feature clusters, and judging the co-occurrence feature modes as incidence relations when the statistical correlation exceeds a preset threshold value to obtain an incidence relation set; S400, constructing a disease development prediction path by adopting a prediction model based on the association relation set, and obtaining trend indexes by traversing path nodes to determine a disease development trend sequence; S500, screening nodes with risk indexes exceeding a preset level from the disease development trend sequence as high-risk nodes, and fusing a plurality of risk factors to calculate potential risk values so as to generate a risk assessment list; s600, generating a comprehensive report according to the risk assessment list, and determining the display sequence of each reference item in the comprehensive report by adopting a priority ordering method; S700, outputting an auxiliary judgment result based on the comprehensive report, and merging new data through a data updating module to dynamically optimize the auxiliary judgment result.
  2. 2. The intelligent medical assistance method based on big data analysis according to claim 1, wherein step S100 comprises: S110, acquiring multi-source medical data transmitted from a heterogeneous medical information terminal, wherein the multi-source medical data comprises structured physiological parameters and unstructured clinical texts; s120, converting the unstructured clinical text into a discrete semantic coding sequence and aligning with the structured physiological parameters to form a multi-modal original data matrix; S130, projecting the multi-mode original data matrix by adopting a dimension alignment matrix to obtain a unified high-dimension feature vector; and S140, performing normalized encoding processing on the unified high-dimensional feature vector to generate a normalized feature representation, and constructing a normalized medical data set based on the normalized feature representation.
  3. 3. The intelligent medical assistance method based on big data analysis according to claim 1, wherein step S200 comprises: S210, acquiring a standardized medical data set, calculating Euclidean distances among features in the standardized medical data set, and determining a feature similarity matrix; s220, constructing a feature density space according to the feature similarity matrix and determining an initial clustering center; s230, performing iterative division on the standardized medical data set based on the initial clustering center by adopting a clustering algorithm to obtain an initial feature group; s240, if the intra-cluster aggregation degree of the initial feature group is larger than a preset aggregation threshold value and the inter-cluster separation degree of the initial feature group is smaller than a preset separation threshold value, executing merging operation to obtain a plurality of feature clusters.
  4. 4. The intelligent medical assistance method based on big data analysis according to claim 3, wherein step S300 comprises: s310, acquiring a source feature index table constructed by a plurality of feature clusters; s320, generating a cross-source feature co-occurrence matrix according to the source feature index table; s330, frequent pattern mining is carried out on the cross-source characteristic co-occurrence matrix, and a cross-data source co-occurrence characteristic pattern is extracted; S340, calculating the statistical correlation of the cross-data source co-occurrence feature mode, and judging the cross-data source co-occurrence feature mode as an association relationship to obtain an association relationship set if the statistical correlation exceeds a preset correlation judgment threshold.
  5. 5. The intelligent medical assistance method based on big data analysis of claim 4, wherein step S400 comprises: s410, extracting clinical representation nodes according to the association relation set; s420, if the state transition probability among the clinical representation nodes is larger than a preset transition threshold, connecting the clinical representation nodes to construct a disease development prediction path; s430, traversing the disease development prediction path to extract disease stage characteristics, and obtaining trend indexes according to the disease stage characteristics; s440, if the time sequence evolution track generated according to the trend index accords with the preset deterioration direction, determining a disease development trend sequence.
  6. 6. The intelligent medical assistance method based on big data analysis according to claim 5, wherein step S500 comprises: S510, extracting abnormal physiological characteristics from the disease development trend sequence to calculate risk indexes; s520, if the risk index is larger than a preset risk level threshold, marking the node corresponding to the risk index as a high risk node; s530, acquiring the deterioration rate corresponding to the high-risk node to extract a plurality of risk factors; S540, carrying out fusion processing on a plurality of risk factors by adopting a weight distribution matrix to obtain potential risk values; S550, performing descending order arrangement according to the potential risk values to generate a risk assessment list.
  7. 7. The intelligent medical assistance method based on big data analysis according to claim 1, wherein step S600 comprises: s610, acquiring potential risk values and pathological description information in a risk assessment list; s620, retrieving a medical knowledge base according to the pathology description information to construct a to-be-sequenced reference item set; S630, quantitatively analyzing the reference item set to be sequenced by utilizing the potential risk value to obtain a severity value and an emergency value; s640, calculating a sorting weight according to the severity value and the emergency value; s650, performing descending order arrangement on the reference item set to be ordered according to the ordering weight so as to obtain an ordered reference item sequence; S660, determining the display sequence of each reference item according to the ordered reference item sequence and generating a comprehensive report, wherein the comprehensive report is generated by packaging the ordered reference item sequence into a structural template.
  8. 8. The intelligent medical assistance method based on big data analysis according to claim 1, wherein step S700 comprises: S710, analyzing an ordered reference item sequence in the comprehensive report, mapping the ordered reference item sequence to a diagnosis and treatment suggestion library, and generating an initial auxiliary judgment result; generating an initial auxiliary judgment result by the following formula: ; Wherein, the In order to assist the judgment result in the initial, For the reference item-diagnosis proposal mapping function, For an ordered sequence of reference items, A diagnosis and treatment suggestion library; S720, collecting a real-time monitoring value associated with the initial auxiliary judgment result and converting the real-time monitoring value into a new data feature vector; S730, re-weighting the initial auxiliary judgment result according to the deviation degree value between the new data feature vector and the reference feature vector to obtain a corrected reference item sequence; S740, packaging the corrected reference item sequence, and outputting an auxiliary judgment result of dynamic optimization.
  9. 9. The intelligent medical assistance method based on big data analysis according to claim 8, wherein in step S730, the deviation degree value between the new data feature vector and the reference feature vector is obtained by the following formula: ; Wherein, the As the value of the degree of deviation, For the feature vector of the new data, Is a reference feature vector; The modified reference term sequence is derived by the following formula: ; Wherein, the For the modified sequence of reference items, Is the deviation influencing coefficient.
  10. 10. A big data analysis based intelligent medical assistance system for performing the big data analysis based intelligent medical assistance method according to any one of claims 1 to 9, comprising: A standardized medical data set acquisition module (10) for acquiring multi-source medical data and converting the multi-source medical data into a unified dimensional feature representation by a data processing module to obtain a standardized medical data set; a plurality of feature cluster acquisition modules (20) for grouping features in the standardized medical dataset using a clustering algorithm to obtain a plurality of feature clusters; The association relation set acquisition module (30) is used for extracting co-occurrence feature modes of cross data sources from a plurality of feature clusters, and judging association relation to obtain an association relation set when the statistical correlation exceeds a preset threshold; The disease development trend sequence determining module (40) is used for constructing a disease development prediction path by adopting a prediction model based on the association relation set, and obtaining trend indexes by traversing path nodes so as to determine a disease development trend sequence; the risk assessment list generation module (50) is used for screening nodes with risk indexes exceeding a preset level from the disease development trend sequence as high-risk nodes, and integrating a plurality of risk factors to calculate potential risk values so as to generate a risk assessment list; a display order determining module (60) for generating a comprehensive report according to the risk assessment list, and determining the display order of each reference item in the comprehensive report by adopting a priority ranking method; And the auxiliary judgment result output and dynamic optimization module (70) is used for outputting an auxiliary judgment result based on the comprehensive report, and integrating new data through the data updating module to dynamically optimize the auxiliary judgment result.

Description

Intelligent medical assistance method and system based on big data analysis Technical Field The invention relates to the technical field of medical treatment, and particularly discloses an intelligent medical treatment assisting method and system based on big data analysis. Background In the medical field, application of informatization and intelligent technology is becoming a key support for improving diagnosis and treatment quality and efficiency. Many medical assistance systems have been tried to support doctor decision making by data, but these methods often face deep integration problems, especially when dealing with information from complex sources, and it is difficult to achieve comprehensive collaboration, so that a doctor still needs to spend a lot of time to spell and judge information in actual diagnosis. One prominent drawback of existing solutions is that they often fail to effectively fuse and mine deeply when faced with medical information of different sources, different formats. For example, electronic medical records, image data and daily patient monitoring data are independent, and lack of a uniform processing mechanism makes it difficult for doctors to quickly obtain comprehensive and consistent reference bases when facing complex diseases. This disadvantage not only increases the workload of the healthcare staff, but also allows the patient to miss the optimal treatment opportunity. Focusing on the technical difficulties, one of the core challenges in the medical field is how to effectively integrate information from multiple channels, forming meaningful auxiliary decisions. The first problem is that the format and content of these information are very different, such as the requirements of text recording and video pictures for storage and analysis are quite different and difficult to handle under the same framework. Further, this discrepancy further makes it difficult to mine out the correlation between information, for example, a test result may be closely related to the long-term life habits of the patient, but due to lack of cross-domain information connection, doctors often cannot perceive the hidden relationship in a short time, and thus the comprehensiveness of diagnosis is affected. Therefore, how to realize unified processing of multiple source information technically and mine hidden associated clues on the basis of the unified processing becomes a key problem for improving the practicability and the accuracy of the medical auxiliary system. Disclosure of Invention The invention provides an intelligent medical auxiliary method and system based on big data analysis, and aims to solve the technical problems that multi-source medical data are difficult to uniformly analyze and cross-source characteristic association cannot be dynamically mined to predict disease development risk. One aspect of the invention relates to an intelligent medical assistance method based on big data analysis, comprising the following steps: S100, acquiring multi-source medical data, and converting the multi-source medical data into a feature representation with uniform dimension through a data processing module to obtain a standardized medical data set; s200, grouping the features in the standardized medical data set by adopting a clustering algorithm to obtain a plurality of feature clusters; S300, extracting co-occurrence feature modes of cross-data sources from a plurality of feature clusters, and judging the co-occurrence feature modes as incidence relations when the statistical correlation exceeds a preset threshold value to obtain an incidence relation set; S400, constructing a disease development prediction path by adopting a prediction model based on the association relation set, and obtaining trend indexes by traversing path nodes to determine a disease development trend sequence; s500, selecting nodes with risk indexes exceeding a preset level from a disease development trend sequence as high-risk nodes, and fusing a plurality of risk factors to calculate potential risk values so as to generate a risk assessment list; s600, generating a comprehensive report according to the risk assessment list, and determining the display sequence of each reference item in the comprehensive report by adopting a priority ranking method; s700, outputting an auxiliary judgment result based on the comprehensive report, and merging new data through a data updating module to dynamically optimize the auxiliary judgment result. Further, step S100 includes: S110, acquiring multi-source medical data transmitted from a heterogeneous medical information terminal, wherein the multi-source medical data comprises structured physiological parameters and unstructured clinical texts; s120, converting unstructured clinical text into a discrete semantic coding sequence and aligning the discrete semantic coding sequence with the structured physiological parameters to form a multi-mode original data matrix; s130, projecting the mult