CN112530582B - Intelligent system for assisting classified coding of death cause - Google Patents
Intelligent system for assisting classified coding of death cause Download PDFInfo
- Publication number
- CN112530582B CN112530582B CN202011439113.3A CN202011439113A CN112530582B CN 112530582 B CN112530582 B CN 112530582B CN 202011439113 A CN202011439113 A CN 202011439113A CN 112530582 B CN112530582 B CN 112530582B
- Authority
- CN
- China
- Prior art keywords
- death
- icd
- text information
- sub
- category
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000012549 training Methods 0.000 claims abstract description 31
- 238000007781 pre-processing Methods 0.000 claims abstract description 16
- 201000010099 disease Diseases 0.000 claims description 38
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 claims description 38
- 238000013527 convolutional neural network Methods 0.000 claims description 27
- 230000001364 causal effect Effects 0.000 claims description 6
- 238000012545 processing Methods 0.000 claims description 6
- 239000013598 vector Substances 0.000 claims description 6
- 238000012795 verification Methods 0.000 claims description 6
- 238000004140 cleaning Methods 0.000 claims description 3
- 238000000605 extraction Methods 0.000 claims description 3
- 239000011159 matrix material Substances 0.000 claims description 3
- 230000011218 segmentation Effects 0.000 claims description 3
- 238000000034 method Methods 0.000 description 5
- 206010028980 Neoplasm Diseases 0.000 description 3
- 208000035473 Communicable disease Diseases 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 208000015181 infectious disease Diseases 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000004891 communication Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 210000000981 epithelium Anatomy 0.000 description 1
- 206010022000 influenza Diseases 0.000 description 1
- 230000003340 mental effect Effects 0.000 description 1
- 238000003058 natural language processing Methods 0.000 description 1
- 230000008520 organization Effects 0.000 description 1
- 238000013441 quality evaluation Methods 0.000 description 1
- 238000007670 refining Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000013519 translation Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/20—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for computer-aided diagnosis, e.g. based on medical expert systems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/35—Clustering; Classification
- G06F16/353—Clustering; Classification into predefined classes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H15/00—ICT specially adapted for medical reports, e.g. generation or transmission thereof
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Medical Informatics (AREA)
- Biomedical Technology (AREA)
- Public Health (AREA)
- General Health & Medical Sciences (AREA)
- Physics & Mathematics (AREA)
- Databases & Information Systems (AREA)
- Epidemiology (AREA)
- Primary Health Care (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Pathology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Evolutionary Computation (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The application provides an intelligent system for assisting in classified coding of death factors, which is characterized by comprising the following components: a data preprocessing module; ICD-10 category training module; the ICD-10 category prediction module predicts the ICD-10 category coding result for the vectorized text information obtained by the data preprocessing module; ICD-10 sub-order text training module; the ICD-10 sub-order text generation module is used for generating direct death factors, intermediate death factors and root death factors in the death factor chain to correspond to ICD-10 sub-order text information according to the vectorized text information obtained through the data preprocessing module; ICD-10 code generation module; ICD-10 code adjustment module. The system provided by the application can automatically generate ICD-10 codes according to the text information in the death report, improves the efficiency of death cause chain analysis, and provides accurate technical support for doctors and medical institutions.
Description
Technical Field
The application relates to an intelligent system for assisting in death factor classification coding, and belongs to the technical fields of disease classification and natural language processing.
Background
In recent years, the domestic medical technology field and the world are more and more frequently communicated, and the unified disease classification standard enables health professionals to exchange health information around the world in a general language. The international classification of diseases (International Classification of Diseases, ICD) is a standard established by the world health organization for representing the classification of diseases by using a coding method, is the basis for the wide communication of medical informatization and medical technology, and ICD-10 is the tenth modification of the prior art. Meanwhile, the disease classification has important guiding significance on multiple aspects of medical teaching scientific research, medical data statistics, medical quality evaluation, medical expense control and the like.
At present, the death cause chain of the death case analysis is manually analyzed and classified according to the ICD-10 standard, and the error classification is very easy to be caused by the subjective factors of classification personnel and the great complexity of task amount. Meanwhile, the current method capable of carrying out auxiliary coding on the death cause classification needs to separately evaluate and analyze category and subgraph class codes of disease classification according to ICD-10 standard, which is time-consuming and labor-consuming, wherein the category represents the general name of a type of disease, such as tumor, infectious disease, mental system disease and the like; subgraph means refining a certain large class of diseases under category, such as epithelial tissue tumor in tumor, influenza in infectious disease, etc.
The existing method is mainly to obtain sub-mesh or detail mesh codes of diseases by comparing and searching from an original ICD-10 standard code library, for example, the application patent CN111210916A mainly codes the diseases conforming to the standard of the code library, obtains classification characteristics of the disease, checks whether the characteristics are matched, and finally outputs the disease. The application patent CN111581987A mainly carries out machine translation and disease classification coding on diagnostic data. The methods involved in the above work are single and lack comprehensive consideration under the corresponding limitation conditions.
Disclosure of Invention
The application aims to solve the technical problems that: the existing death cause classification method is mainly used for obtaining sub-mesh or detail mesh codes of diseases by comparing and searching from an original ICD-10 standard code library, is single and lacks comprehensive consideration under corresponding limiting conditions.
In order to solve the technical problems, the technical scheme of the application provides an intelligent system for assisting death factor classification coding, which is characterized by comprising the following steps:
the data preprocessing module is used for vectorizing the text information of the input death report;
ICD-10 category training module, build vectorization text information data set and train Multi-Task CNN category predictive model;
the ICD-10 category prediction module is used for classifying the vectorized text information obtained through the data preprocessing module by using a Multi-Task CNN category prediction model to predict direct death factors, intermediate death factors and root death factors in the death factor chain to correspond to ICD-10 category coding results;
the ICD-10 sub-order text training module is used for constructing a vectorized text information data set and training a seq2seq sub-order text generation model;
the ICD-10 sub-order text generation module is used for generating direct death factors, intermediate death factors and root death factors in the death factor chain according to the vectorized text information obtained through the data preprocessing module and corresponding to the ICD-10 sub-order text information;
the ICD-10 code generating module is used for searching in a disease library corresponding to the ICD-10 category code result generated in the ICD-10 category predicting module according to the ICD-10 subgraph text information to acquire an ICD-10 code of the current death report;
and the ICD-10 coding adjustment module is used for adding and deleting the result of the death cause chain or adjusting the causal sequence of the death cause chain.
Preferably, the implementation of the data preprocessing module includes the following steps:
step S101: inputting text information in a death report;
step S102: segmenting the text information input in the step S101, and obtaining segmented text information according to a specified standard word stock;
step S103: data cleaning is carried out on the text information after word segmentation, useless labels, special symbols, stop words and the like are removed, and cleaned text information is obtained;
step S104: performing standardization processing on the text information obtained in the step S103, and obtaining standardized text information according to the corresponding word stock;
step S105: performing vectorization processing on the text information obtained in the step S104, and performing word embedding coding to obtain a vectorized text information queue;
step S106: and sending the vectorized text information queue to an ICD-10 category prediction module and an ICD-10 sub-category text generation module.
Preferably, the implementation of the ICD-10 category training module includes the steps of:
step S601: constructing a data set for training a Multi-Task CNN category prediction model by using the vectorized text information queue;
step S602: inputting the text vectors in the data set obtained in the step S601 and the corresponding 26 category labels into a Multi-Task CNN category prediction model;
step S603: and obtaining a Multi-Task CNN category prediction model used in the ICD-10 category prediction module until the loss values and the classification accuracy of the training set and the verification set samples reach set thresholds.
Preferably, the implementation of the ICD-10 category prediction module includes the steps of:
step S201: inputting a vectorized text information queue to be predicted into a Multi-Task CNN category prediction model;
step S202: the Multi-Task CNN category prediction model utilizes a convolutional neural network to provide the characteristics of a quantized text information queue;
step S203: the Multi-Task CNN category prediction model utilizes a classifier to respectively generate ICD-10 category coding results of direct death factors, intermediate death factors and root death factors in the death factor chain according to the characteristics obtained in the step S202.
Preferably, the implementation of the ICD-10 sub-order text training module comprises the following steps:
step S701: constructing a dataset for training a seq2seq sub-order text generation model using the vectorized text information queue;
step S702: inputting the text vector queue and the corresponding sub-view text label in the dataset obtained in the step S701 into a seq2seq sub-view text generation model;
step S703: until the loss values and the prediction accuracy of the training set and the verification set samples reach set thresholds, acquiring a seq2seq sub-order text generation model used in the ICD-10 sub-order text generation module.
Preferably, the implementation of the ICD-10 sub-order text generation module comprises the following steps:
step S301: inputting a vectorized text information queue to be predicted into a seq2seq sub-order text generation model;
step S302: feature extraction, namely calculating the features of the text information according to the weights of the coding network by using a seq2seq sub-order text generation model to obtain a hidden feature queue of the vectorized text information;
step S303: attention adjustment, namely multiplying the feature queue by an attention matrix to acquire attention feature queues containing different attention weight information;
step S304: and decoding the network according to the attention characteristic queue to generate sub-order texts of the direct death cause, the intermediate death cause and the root death cause in the death cause chain, and acquiring ICD-10 sub-order text information.
Preferably, the implementation of the ICD-10 code generation module includes the following steps:
step S401: inputting an ICD-10 category encoding result generated by the ICD-10 category prediction module and ICD-10 sub-category text information generated by the ICD-10 sub-category text generation module;
step S402: according to ICD-10 category coding results, three death factor disease libraries of direct death factor, intermediate death factor and root death factor in corresponding categories and corresponding disease subgraph codes are obtained;
step S403: according to the generated ICD-10 sub-order text information, searching the three types of death factor libraries for the disease with the largest matching degree in the step S402, and obtaining ICD-10 sub-order codes and disease names thereof corresponding to the direct death factor, the intermediate death factor and the root death factor, wherein the direct death factor is the cause of death of a patient, the intermediate death factor is the cause of the direct death factor, the root death factor is the cause of the intermediate death factor, and the direct death factor, the intermediate death factor and the root death factor jointly form a death factor chain of the patient.
Preferably, the implementation of the ICD-10 coding adjustment module includes the following steps:
step S501: inputting ICD-10 subgraph codes of direct death cause, intermediate death cause and root death cause and disease names thereof;
step S502: selecting whether to add, delete or modify direct and intermediate and root causes in the cause chain;
step S503: and selecting whether to adjust the causal sequence of the direct death factor, the intermediate death factor and the root death factor in the death factor chain, and acquiring the final death factor chain ICD-10 code and the disease name thereof.
The system provided by the application can automatically generate ICD-10 codes and disease names thereof according to text information in the death report, improves the efficiency of death chain analysis, and provides accurate technical support for doctors and medical institutions.
Drawings
FIG. 1 is a schematic diagram of an intelligent system for assisting in classified encoding of death factors;
FIG. 2 is a flow chart of a data preprocessing module according to the present application;
FIG. 3 is a flow chart of the ICD-10 training module of the present application;
FIG. 4 is a flow chart of an ICD-10 category prediction module of the present application;
FIG. 5 is a flow chart of the ICD-10 sub-order text training module of the present application;
FIG. 6 is a flow chart of the ICD-10 sub-order text generation module of the present application;
FIG. 7 is a flow chart of the ICD-10 code generation module of the present application;
fig. 8 is a flow chart of the ICD-10 code adjustment module of the present application.
Detailed Description
The application will be further illustrated with reference to specific examples. It is to be understood that these examples are illustrative of the present application and are not intended to limit the scope of the present application. Furthermore, it should be understood that various changes and modifications can be made by one skilled in the art after reading the teachings of the present application, and such equivalents are intended to fall within the scope of the application as defined in the appended claims.
As shown in fig. 1, the intelligent system for assisting in classified encoding of death causes provided by the application comprises the following contents:
the data preprocessing module is used for vectorizing the text information of the input death report;
ICD-10 category training module, build vectorization text information data set and train Multi-Task CNN category predictive model;
the ICD-10 category prediction module is used for classifying the vectorized text information obtained through the data preprocessing module by using a Multi-Task CNN category prediction model to predict direct death factors, intermediate death factors and root death factors in the death factor chain to correspond to ICD-10 category coding results;
the ICD-10 sub-order text training module is used for constructing a vectorized text information data set and training a seq2seq sub-order text generation model;
the ICD-10 sub-order text generation module is used for generating direct death factors, intermediate death factors and root death factors in the death factor chain according to the vectorized text information obtained through the data preprocessing module and corresponding to the ICD-10 sub-order text information;
the ICD-10 code generating module is used for searching in a disease library corresponding to the ICD-10 category code result generated in the ICD-10 category predicting module according to the ICD-10 subgraph text information to acquire an ICD-10 code of the current death report;
and the ICD-10 coding adjustment module is used for adding and deleting the result of the death cause chain or adjusting the causal sequence of the death cause chain.
The data preprocessing module, as shown in fig. 2, comprises the following steps:
step S101: inputting text information in a death report;
step S102: segmenting the text information input in the step S101, and obtaining segmented text information according to a specified standard word stock;
step S103: data cleaning is carried out on the text information after word segmentation, useless labels, special symbols, stop words and the like are removed, and cleaned text information is obtained;
step S104: performing standardization processing on the text information obtained in the step S103, and obtaining standardized text information according to the corresponding word stock;
step S105: performing vectorization processing on the text information obtained in the step S104, and performing word embedding coding to obtain a vectorized text information queue;
step S106: and sending the vectorized text information queue to an ICD-10 category prediction module and an ICD-10 sub-category text generation module.
ICD-10 category training module, as shown in FIG. 3, is implemented by:
step S601: constructing a data set for training a Multi-Task CNN category prediction model by using the vectorized text information queue;
step S602: inputting the text vectors in the data set obtained in the step S601 and the corresponding 26 category labels into a Multi-Task CNN category prediction model;
step S603: and obtaining a Multi-Task CNN category prediction model used in the ICD-10 category prediction module until the loss values and the classification accuracy of the training set and the verification set samples reach set thresholds.
ICD-10 category prediction module, as shown in FIG. 4, is implemented by:
step S201: inputting a vectorized text information queue to be predicted into a Multi-Task CNN category prediction model;
step S202: the Multi-Task CNN category prediction model utilizes a convolutional neural network to provide the characteristics of a quantized text information queue;
step S203: the Multi-Task CNN category prediction model utilizes a classifier to respectively generate ICD-10 category coding results of direct death factors, intermediate death factors and root death factors in the death factor chain according to the characteristics obtained in the step S202.
The ICD-10 sub-order text training module, as shown in FIG. 5, is implemented by the following steps:
step S701: constructing a dataset for training a seq2seq sub-order text generation model using the vectorized text information queue;
step S702: inputting the text vector queue and the corresponding sub-view text label in the dataset obtained in the step S701 into a seq2seq sub-view text generation model;
step S703: until the loss values and the prediction accuracy of the training set and the verification set samples reach set thresholds, acquiring a seq2seq sub-order text generation model used in the ICD-10 sub-order text generation module.
The ICD-10 sub-order text generation module, as shown in FIG. 6, is implemented by:
step S301: inputting a vectorized text information queue to be predicted into a seq2seq sub-order text generation model;
step S302: feature extraction, namely calculating the features of the text information according to the weights of the coding network by using a seq2seq sub-order text generation model to obtain a hidden feature queue of the vectorized text information;
step S303: attention adjustment, namely multiplying the feature queue by an attention matrix to acquire attention feature queues containing different attention weight information;
step S304: and decoding the network according to the attention characteristic queue to generate sub-order texts of the direct death cause, the intermediate death cause and the root death cause in the death cause chain, and acquiring ICD-10 sub-order text information.
The ICD-10 code generation module, as shown in FIG. 7, comprises the following steps:
step S401: inputting an ICD-10 category encoding result generated by the ICD-10 category prediction module and ICD-10 sub-category text information generated by the ICD-10 sub-category text generation module;
step S402: according to ICD-10 category coding results, three death factor disease libraries of direct death factor, intermediate death factor and root death factor in corresponding categories and corresponding disease subgraph codes are obtained;
step S403: according to the generated ICD-10 sub-order text information, searching the three types of death factor libraries for the disease with the largest matching degree in the step S402, and obtaining ICD-10 sub-order codes and disease names thereof corresponding to the direct death factor, the intermediate death factor and the root death factor, wherein the direct death factor is the cause of death of a patient, the intermediate death factor is the cause of the direct death factor, the root death factor is the cause of the intermediate death factor, and the direct death factor, the intermediate death factor and the root death factor jointly form a death factor chain of the patient.
The ICD-10 code adjustment module, as shown in FIG. 8, comprises the following steps:
step S501: inputting ICD-10 subgraph codes of direct death cause, intermediate death cause and root death cause and disease names thereof;
step S502: selecting whether to add, delete or modify direct and intermediate and root causes in the cause chain;
step S503: and selecting whether to adjust the causal sequence of the direct cause, the intermediate cause and the root cause in the cause chain by dragging, and acquiring the final cause chain ICD-10 code and the disease name thereof.
Claims (7)
1. An intelligent system for assisting in the classified encoding of a cause of death, comprising:
the data preprocessing module is used for vectorizing the text information of the input death report;
ICD-10 category training module, build vectorization text information data set and train Multi-Task CNN category predictive model;
the ICD-10 category prediction module is used for classifying the vectorized text information obtained through the data preprocessing module by using a Multi-Task CNN category prediction model to predict direct death factors, intermediate death factors and root death factors in the death factor chain to correspond to ICD-10 category coding results;
the ICD-10 sub-order text training module is used for constructing a vectorized text information data set and training a seq2seq sub-order text generation model;
the ICD-10 sub-order text generation module is used for generating direct death factors, intermediate death factors and root death factors in the death factor chain according to the vectorized text information obtained through the data preprocessing module and corresponding to the ICD-10 sub-order text information;
the ICD-10 code generating module is used for searching in a disease library corresponding to the ICD-10 category code result generated in the ICD-10 category predicting module according to the ICD-10 subgraph text information to acquire an ICD-10 code of the current death report; the ICD-10 code generation module comprises the following steps:
step S401: inputting an ICD-10 category encoding result generated by the ICD-10 category prediction module and ICD-10 sub-category text information generated by the ICD-10 sub-category text generation module;
step S402: according to ICD-10 category coding results, three death factor disease libraries of direct death factor, intermediate death factor and root death factor in corresponding categories and corresponding disease subgraph codes are obtained;
step S403: searching the three types of death factor libraries for the disease with the largest matching degree according to the generated ICD-10 sub-order text information in the step S402, and obtaining ICD-10 sub-order codes and disease names thereof corresponding to the direct death factor, the intermediate death factor and the root cause, wherein the direct death factor is the cause of death of a patient, the intermediate death factor is the cause of the direct death factor, the root cause is the cause of the intermediate death factor, and the direct death factor, the intermediate death factor and the root cause jointly form a death factor chain of the patient;
and the ICD-10 coding adjustment module is used for adding and deleting the result of the death cause chain or adjusting the causal sequence of the death cause chain.
2. An intelligent system for assisting in the classification encoding of causes of death according to claim 1, wherein the implementation of the data preprocessing module comprises the steps of:
step S101: inputting text information in a death report;
step S102: segmenting the text information input in the step S101, and obtaining segmented text information according to a specified standard word stock;
step S103: data cleaning is carried out on the text information after word segmentation, useless labels, special symbols, stop words and the like are removed, and cleaned text information is obtained;
step S104: performing standardization processing on the text information obtained in the step S103, and obtaining standardized text information according to the corresponding word stock;
step S105: performing vectorization processing on the text information obtained in the step S104, and performing word embedding coding to obtain a vectorized text information queue;
step S106: and sending the vectorized text information queue to an ICD-10 category prediction module and an ICD-10 sub-category text generation module.
3. An intelligent system for assisting in the classification encoding of causes of death according to claim 1, wherein the implementation of said ICD-10 category training module comprises the steps of:
step S601: constructing a data set for training a Multi-Task CNN category prediction model by using the vectorized text information queue;
step S602: inputting the text vectors in the data set obtained in the step S601 and the corresponding 26 category labels into a Multi-Task CNN category prediction model;
step S603: and obtaining a Multi-Task CNN category prediction model used in the ICD-10 category prediction module until the loss values and the classification accuracy of the training set and the verification set samples reach set thresholds.
4. An intelligent system for assisting in the classification encoding of causes of death according to claim 1, wherein said implementation of said ICD-10 category prediction module comprises the steps of:
step S201: inputting a vectorized text information queue to be predicted into a Multi-Task CNN category prediction model;
step S202: the Multi-Task CNN category prediction model utilizes a convolutional neural network to provide the characteristics of a quantized text information queue;
step S203: the Multi-Task CNN category prediction model utilizes a classifier to respectively generate ICD-10 category coding results of direct death factors, intermediate death factors and root death factors in the death factor chain according to the characteristics obtained in the step S202.
5. An intelligent system for assisting in the classification encoding of causes of death according to claim 1, wherein the implementation of said ICD-10 sub-order text training module comprises the steps of:
step S701: constructing a dataset for training a seq2seq sub-order text generation model using the vectorized text information queue;
step S702: inputting the text vector queue and the corresponding sub-view text label in the dataset obtained in the step S701 into a seq2seq sub-view text generation model;
step S703: until the loss values and the prediction accuracy of the training set and the verification set samples reach set thresholds, acquiring a seq2seq sub-order text generation model used in the ICD-10 sub-order text generation module.
6. An intelligent system for assisting in the encoding of a death cause classification according to claim 1, wherein said ICD-10 sub-order text generation module is implemented by:
step S301: inputting a vectorized text information queue to be predicted into a seq2seq sub-order text generation model;
step S302: feature extraction, namely calculating the features of the text information according to the weights of the coding network by using a seq2seq sub-order text generation model to obtain a hidden feature queue of the vectorized text information;
step S303: attention adjustment, namely multiplying the feature queue by an attention matrix to acquire attention feature queues containing different attention weight information;
step S304: and decoding the network according to the attention characteristic queue to generate sub-order texts of the direct death cause, the intermediate death cause and the root death cause in the death cause chain, and acquiring ICD-10 sub-order text information.
7. An intelligent system for assisting in the classification coding of causes of death according to claim 1, wherein said ICD-10 code adjustment module is implemented by:
step S501: inputting ICD-10 subgraph codes of direct death cause, intermediate death cause and root death cause and disease names thereof;
step S502: selecting whether to add, delete or modify direct and intermediate and root causes in the cause chain;
step S503: and selecting whether to adjust the causal sequence of the direct death factor, the intermediate death factor and the root death factor in the death factor chain, and acquiring the final death factor chain ICD-10 code and the disease name thereof.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011439113.3A CN112530582B (en) | 2020-12-11 | 2020-12-11 | Intelligent system for assisting classified coding of death cause |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011439113.3A CN112530582B (en) | 2020-12-11 | 2020-12-11 | Intelligent system for assisting classified coding of death cause |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112530582A CN112530582A (en) | 2021-03-19 |
| CN112530582B true CN112530582B (en) | 2023-11-14 |
Family
ID=74999465
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011439113.3A Active CN112530582B (en) | 2020-12-11 | 2020-12-11 | Intelligent system for assisting classified coding of death cause |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112530582B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11515017B2 (en) * | 2021-02-01 | 2022-11-29 | Konica Minolta Healthcare Americas, Inc. | Deletion of medical information shared between management server and healthcare facilities |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107545143A (en) * | 2017-09-04 | 2018-01-05 | 复旦大学 | The mapping method of disease and human body |
| CN109299273A (en) * | 2018-11-02 | 2019-02-01 | 广州语义科技有限公司 | Based on the multi-source multi-tag file classification method and its system for improving seq2seq model |
| CN109994216A (en) * | 2019-03-21 | 2019-07-09 | 上海市第六人民医院 | A kind of ICD intelligent diagnostics coding method based on machine learning |
| CN111210916A (en) * | 2019-12-23 | 2020-05-29 | 望海康信(北京)科技股份公司 | Medical record home page coding method and system |
| CN111462896A (en) * | 2020-03-31 | 2020-07-28 | 重庆大学 | Real-time intelligent auxiliary ICD coding system and method based on medical record |
| CN111540468A (en) * | 2020-04-21 | 2020-08-14 | 重庆大学 | ICD automatic coding method and system for visualization of diagnosis reason |
| CN112599213A (en) * | 2021-03-04 | 2021-04-02 | 联仁健康医疗大数据科技股份有限公司 | Classification code determining method, device, equipment and storage medium |
| CN113988013A (en) * | 2021-10-25 | 2022-01-28 | 清华大学 | ICD coding method and device based on multitask learning and graph attention network |
| CN115270718A (en) * | 2022-07-26 | 2022-11-01 | 中国医学科学院阜外医院 | Automatic cataloging method and system for disease codes |
| CN116013546A (en) * | 2022-12-22 | 2023-04-25 | 联通(广东)产业互联网有限公司 | Death factor chain detection model training and detection method based on disease Chinese interpretation |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR101410230B1 (en) * | 2007-08-17 | 2014-06-20 | 삼성전자주식회사 | Audio encoding method and apparatus, and audio decoding method and apparatus, processing death sinusoid and general continuation sinusoid in different way |
-
2020
- 2020-12-11 CN CN202011439113.3A patent/CN112530582B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107545143A (en) * | 2017-09-04 | 2018-01-05 | 复旦大学 | The mapping method of disease and human body |
| CN109299273A (en) * | 2018-11-02 | 2019-02-01 | 广州语义科技有限公司 | Based on the multi-source multi-tag file classification method and its system for improving seq2seq model |
| CN109994216A (en) * | 2019-03-21 | 2019-07-09 | 上海市第六人民医院 | A kind of ICD intelligent diagnostics coding method based on machine learning |
| CN111210916A (en) * | 2019-12-23 | 2020-05-29 | 望海康信(北京)科技股份公司 | Medical record home page coding method and system |
| CN111462896A (en) * | 2020-03-31 | 2020-07-28 | 重庆大学 | Real-time intelligent auxiliary ICD coding system and method based on medical record |
| CN111540468A (en) * | 2020-04-21 | 2020-08-14 | 重庆大学 | ICD automatic coding method and system for visualization of diagnosis reason |
| CN112599213A (en) * | 2021-03-04 | 2021-04-02 | 联仁健康医疗大数据科技股份有限公司 | Classification code determining method, device, equipment and storage medium |
| CN113988013A (en) * | 2021-10-25 | 2022-01-28 | 清华大学 | ICD coding method and device based on multitask learning and graph attention network |
| CN115270718A (en) * | 2022-07-26 | 2022-11-01 | 中国医学科学院阜外医院 | Automatic cataloging method and system for disease codes |
| CN116013546A (en) * | 2022-12-22 | 2023-04-25 | 联通(广东)产业互联网有限公司 | Death factor chain detection model training and detection method based on disease Chinese interpretation |
Non-Patent Citations (10)
| Title |
|---|
| 《死因编码与疾病诊断编码的鉴别分析》;吴晓云;《中国病案》;第12卷(第6期);32-33 * |
| 2010年佛山市医疗机构死因信息登记质量分析;李艳容;黄嘉殷;古冠斌;;中外医学研究(第16期);全文 * |
| ICD-10合并编码问题的研究与实现;孙晶;《医药卫生科技》(第4期);1-90 * |
| 吴晓云 ; 谭礼萍 ; 俞长荣 ; 郑银雄 ; .死因编码与疾病诊断编码的鉴别分析.中国病案.2011,(第06期),全文. * |
| 姬一兵 ; 王黎君 ; 周脉耕 ; .根本死因自动编码工具在死因监测工作中的编码实例分析.中华疾病控制杂志.2013,(第09期),全文. * |
| 徐海英 ; .死因诊断实例分析.中国病案.2012,(第06期),全文. * |
| 李艳容 ; 黄嘉殷 ; 古冠斌 ; .2010年佛山市医疗机构死因信息登记质量分析.中外医学研究.2013,(第16期),全文. * |
| 根本死因自动编码工具在死因监测工作中的编码实例分析;姬一兵;王黎君;周脉耕;;中华疾病控制杂志(第09期);全文 * |
| 死因编码与疾病诊断编码的鉴别分析;吴晓云;谭礼萍;俞长荣;郑银雄;;中国病案(第06期);全文 * |
| 死因诊断实例分析;徐海英;;中国病案(第06期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112530582A (en) | 2021-03-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109299472B (en) | Text data processing method and device, electronic equipment and computer readable medium | |
| CN111222340B (en) | Breast electronic medical record entity recognition system based on multi-standard active learning | |
| CN111177186B (en) | Single sentence intention recognition method, device and system based on question retrieval | |
| CN113191148B (en) | Rail transit entity identification method based on semi-supervised learning and clustering | |
| CN110580308A (en) | information auditing method and device, electronic equipment and storage medium | |
| CN112163420A (en) | NLP technology-based RPA process automatic generation method | |
| CN111274817A (en) | Intelligent software cost measurement method based on natural language processing technology | |
| CN114880468A (en) | Building specification examination method and system based on BilSTM and knowledge graph | |
| CN114021584A (en) | Knowledge representation learning method based on graph convolution network and translation model | |
| CN115359799A (en) | Speech recognition method, training method, device, electronic equipment and storage medium | |
| CN116070632A (en) | Informal text entity tag identification method and device | |
| CN112530582B (en) | Intelligent system for assisting classified coding of death cause | |
| CN111985680A (en) | Criminal multi-criminal name prediction method based on capsule network and time sequence | |
| CN113761875B (en) | Event extraction method and device, electronic equipment and storage medium | |
| CN114356924A (en) | Method and apparatus for extracting data from structured documents | |
| CN112579777B (en) | Semi-supervised classification method for unlabeled text | |
| CN110232328A (en) | A kind of reference report analytic method, device and computer readable storage medium | |
| CN117422074A (en) | Method, device, equipment and medium for standardizing clinical information text | |
| CN114388108A (en) | User feedback analysis method based on multi-task learning | |
| CN115408506B (en) | NL2SQL method combining semantic analysis and semantic component matching | |
| CN116362247A (en) | Entity extraction method based on MRC framework | |
| CN116186422A (en) | Disease-related public opinion analysis system based on social media and artificial intelligence | |
| CN114611489A (en) | Text logic condition extraction AI model construction method, extraction method and system | |
| CN111353291B (en) | Method and system for calculating optimal annotation set based on complaint work order training text | |
| CN116933750B (en) | Instrument price inquiring information extracting method, device and storage medium based on semantic similarity and knowledge base |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |