CN109670179A - Case history text based on iteration expansion convolutional neural networks names entity recognition method - Google Patents
Case history text based on iteration expansion convolutional neural networks names entity recognition method Download PDFInfo
- Publication number
- CN109670179A CN109670179A CN201811563980.0A CN201811563980A CN109670179A CN 109670179 A CN109670179 A CN 109670179A CN 201811563980 A CN201811563980 A CN 201811563980A CN 109670179 A CN109670179 A CN 109670179A
- Authority
- CN
- China
- Prior art keywords
- label
- neural networks
- convolutional neural
- word
- model
- 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.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
- G06F40/289—Phrasal analysis, e.g. finite state techniques or chunking
- G06F40/295—Named entity recognition
-
- 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
- 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
- G06N3/047—Probabilistic or stochastic networks
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- General Physics & Mathematics (AREA)
- Computational Linguistics (AREA)
- General Engineering & Computer Science (AREA)
- Biomedical Technology (AREA)
- Evolutionary Computation (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Data Mining & Analysis (AREA)
- Biophysics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- Probability & Statistics with Applications (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Medical Treatment And Welfare Office Work (AREA)
- Character Discrimination (AREA)
Abstract
The present invention provides a kind of case history text name entity recognition method based on iteration expansion convolutional neural networks, this method is named Entity recognition in medical electronics medical record data collection CCKS2017, input one section of Chinese electronic health record text, use iteration expansion convolutional neural networks and condition random field as model framework, use Chinese radical as feature, to extract the name entity in text, such as disease name, detection methods.
Description
Technical field
The present invention relates to natural language processings and clinical relevant fields, more particularly, to one kind based on iteration expansion volume
The case history text of product neural network names entity recognition method.
Background technique
In recent years, with the development of big data and computer technology, more and more medical institutions start using electronics disease
Go through system.Electronic medical record system is medicine special-purpose software.Hospital records patient assessment's by electronic health record in a manner of electronic
Information, comprising: medical history, checks inspection result, doctor's advice, operation record, nursing record etc. at progress note, wherein existing structure
Change information, also has non-structured free text, there are also Figure and Images.
With the development of artificial intelligence technology, many team begin trying artificial intelligence technology being used for medical field, with
As a kind of medical assistance means.Electronic health record is as a kind of important medical data, and it comprises many non-structured texts.
Analysis to unstructured case history text is to allow computer understanding case history, the basis using case history.Based on the structure to case history
Change, the relationship and its probability between multiple knowledge points such as symptom, disease, drug, inspection inspection can be calculated, construct medical treatment neck
The knowledge mapping in domain advanced optimizes the work of doctor.
The structuring of case history text, a kind of important means name Entity recognition.I.e. given one section of medical text, is extracted
The medicine entity of specified type out, and they are referred in the classification pre-defined, classification include symptom, physical feeling,
Treatment, disease, inspection item etc..Such as: " for patient by complex treatment, neck-shoulder pain symptom is substantially reduced ", traditional Chinese medicine is real
Body includes " shoulder neck " (physical feeling), " pain " (symptom).
The name Entity recognition of medical domain is different from general domain, and the main distinction is as follows: (1) many of medical field
Professional term and rarely used word, such as " loratadine tablet ", current Chinese word segmentation tool cannot be segmented well, thus can shadow
Ring subsequent recognition effect.(2) part entity title is longer, such as " Cerebrolysin Vial nourishing brain cell " (treatment), part mould
Type has been difficult to set up longer Context-dependent.
For first problem, it is contemplated that participle effect of the existing participle tool on medicine text is poor, herein I
No longer segmented, directly Chinese character is operated.On the one hand other portions of erroneous effects model caused by being avoided that because of participle
Point, model vocabulary size is on the other hand also reduced, parameter is reduced, avoids over-fitting.In addition, for case history text, it may appear that big
Amount has the character of specific radical, such as chest, liver, spleen, lung human organ are all by " moon " word, in addition " cancer, acute diseases such as cholera and sunstroke, hemorrhoid, phlegm " etc. with
" Epileptic " is the word of radical, all related to disease or symptom, therefore we are input to radical as feature in model, to alleviate life
The problems such as rare word.For Second Problem, it is contemplated that enabling model to read in long range using expansion convolutional neural networks
Hereafter, too big without regard to convolution kernel is made.To sum up, we have proposed based on expansion convolutional neural networks and Chinese radical feature
Case history name entity recognition method.
Summary of the invention
The present invention provides a kind of case history that convolutional neural networks are expanded for extracting the name entity in text based on iteration
Text names entity recognition method.
In order to reach above-mentioned technical effect, technical scheme is as follows:
A kind of case history text name entity recognition method based on iteration expansion convolutional neural networks, comprising the following steps:
S1: the model for naming the iteration of Entity recognition to expand convolutional neural networks and condition random field is established;
S2: the loss function of model is established;
S3: the training of model is carried out, and is tested on test set.
Further, the detailed process of the step S1 is:
S11: building Embedding, since model will handle text, and text cannot directly be handled by model, need elder generation
Text conversion is indicated at vector, i.e., with Embedding layers of completion, the vector including word is indicated and the vector of its radical indicates;
S12: building iteration expands convolutional neural networks, and for extracting feature, expansion convolutional neural networks include four layers swollen
Swollen convolutional layer, expansion radius are respectively 1,2,3,3, and every layer includes 100 convolution kernels, and each convolution kernel width is 3, the last layer
Output be re-entered into first layer, i.e. our so-called iteration, iteration 4 times altogether;
S13: building conditional random field models, the feature that previous step is extracted is as the input of condition random field, condition
Random field exports a sequence label to each word, to mark whether the word is a part of entity, if it does, being entity
Beginning, centre or end up, and belong to what kind of entity.
Further, the detailed process of the step S2 is:
S21: loss function is provided by negative log-likelihood function, and likelihood function is equal to, and predicts the score of the label come,
Than the score of upper all possible label:
Wherein, s (x, y) is score function;
S22: the calculating of score is divided into two parts, conversion score A i, j between (1) label are transformed into mark from label i
Sign the score of j;(2) label score Pm, the n of word, that is, give some word m, and label is the score of n, it may be assumed that
Further, detailed process is as follows by the step S3:
S31: to input text, splitting individual character processing, and each word obtains its radical, by Embedding layer acquisition word with
The vector of radical indicates, is input to iteration expansion convolutional neural networks to extract feature and the feature extracted is input to condition
In random field, final label is obtained;
S32: comparing label and the model answer of prediction, and loss function is calculated by mode described in step S2, uses
Adam optimizer, Lai Youhua loss function update model parameter;
S33: by data set according to ratio cut partition training set, the test set of 9:1, repetitive exercise 100 times on training set,
Penalty values have tended towards stability after 50 times, the model after training are saved, and the test result on test set.It is right when test result
Each word of test sample all exports a corresponding label;
S34: S31-S33 is repeated, 5 cross validations are done on test set, use the indexs such as accurate rate, recall rate, F1 value
The effect of model is measured, takes 5 average values as last effect.
Compared with prior art, the beneficial effect of technical solution of the present invention is:
The present invention provides a kind of method of case history text name Entity recognition based on iteration expansion convolutional neural networks, this
Invention is named Entity recognition in medical electronics medical record data collection CCKS2017, inputs one section of Chinese electronic health record text, uses
Iteration expands convolutional neural networks and condition random field as model framework, uses Chinese radical as feature, to extract in text
Name entity, such as disease name, detection methods.
Detailed description of the invention
Fig. 1 is flow diagram of the present invention;
Fig. 2 is the algorithm structure schematic diagram in embodiment 1.
Specific embodiment
The attached figures are only used for illustrative purposes and cannot be understood as limitating the patent;
In order to better illustrate this embodiment, the certain components of attached drawing have omission, zoom in or out, and do not represent actual product
Size;
To those skilled in the art, it is to be understood that certain known features and its explanation, which may be omitted, in attached drawing
's.
The following further describes the technical solution of the present invention with reference to the accompanying drawings and examples.
Embodiment 1
As shown in Figure 1, a kind of case history text based on iteration expansion convolutional neural networks names entity recognition method, including
Following steps:
S1: the model for naming the iteration of Entity recognition to expand convolutional neural networks and condition random field is established;
S2: the loss function of model is established;
S3: the training of model is carried out, and is tested on test set.
The detailed process of step S1 is:
S11: building Embedding, since model will handle text, and text cannot directly be handled by model, need elder generation
Text conversion is indicated at vector, i.e., with Embedding layers of completion, the vector including word is indicated and the vector of its radical indicates;
S12: building iteration expands convolutional neural networks, and for extracting feature, expansion convolutional neural networks include four layers swollen
Swollen convolutional layer, expansion radius are respectively 1,2,3,3, and every layer includes 100 convolution kernels, and each convolution kernel width is 3, the last layer
Output be re-entered into first layer, i.e. our so-called iteration, iteration 4 times altogether;
S13: building conditional random field models, the feature that previous step is extracted is as the input of condition random field, condition
Random field exports a sequence label to each word, to mark whether the word is a part of entity, if it does, being entity
Beginning, centre or end up, and belong to what kind of entity.
The detailed process of step S2 is:
S21: loss function is provided by negative log-likelihood function, and likelihood function is equal to, and predicts the score of the label come,
Than the score of upper all possible label:
Wherein, s (x, y) is score function;
S22: the calculating of score is divided into two parts, conversion score A i, j between (1) label are transformed into mark from label i
Sign the score of j;(2) label score Pm, the n of word, that is, give some word m, and label is the score of n, it may be assumed that
Detailed process is as follows by step S3:
S31: to input text, splitting individual character processing, and each word obtains its radical, by Embedding layer acquisition word with
The vector of radical indicates, is input to iteration expansion convolutional neural networks to extract feature and the feature extracted is input to condition
In random field, final label is obtained;
S32: comparing label and the model answer of prediction, and loss function is calculated by mode described in step S2, uses
Adam optimizer, Lai Youhua loss function update model parameter;
S33: by data set according to ratio cut partition training set, the test set of 9:1, repetitive exercise 100 times on training set,
Penalty values have tended towards stability after 50 times, the model after training are saved, and the test result on test set.It is right when test result
Each word of test sample all exports a corresponding label;
S34: S31-S33 is repeated, 5 cross validations are done on test set, use the indexs such as accurate rate, recall rate, F1 value
The effect of model is measured, takes 5 average values as last effect.
The present invention be directed to the name Entity recognition of case history text, the data set that we use is CCKS2017 Chinese electronics
Case history Entity recognition data set is issued by national knowledge mapping and semantic computation conference.Data set owner will be contained to be led with medicine
The relevant text in domain.The entity class being related to includes: organ (body), symptom (symptom), checks (check), disease
Disease (disease), treatment method (treatment), each entity distribution is as shown in table 1 in data set.Data set notation methods are
BIOES, i.e. Begin (entity starts), Intermediate (among entity), End (entity ending), (single word is Single
One entity), O (non-physical)." B-body, E-body, B-symptom, I- are labeled as such as " abdominal pain sense disappearance "
Symptom, E-symptom, O, O ", then we from this mark it is found that " abdomen " be type for " organ " entity, and
" feeling of pain " is the entity that type is " symptom ", and " disappearance " is not entity.
The distribution situation of table 1, training set entity
In existing method, the preferable way of effect be in conjunction with term vector two-way shot and long term memory network (LSTM)+
Condition random field (CRF).Wherein shot and long term memory network is for understanding read statement and extracting feature, and condition random field is for producing
Raw label.But in case history text, speech habits are different from general term, generally simplify, rigorously, and context relation is not
Greatly, therefore shot and long term memory network herein and is not suitable for.In addition, term vector is also not suitable for herein, because exist in case history text compared with
Multi-specialized noun, has segmented very big difficulty to it, and mistake caused by segmenting can accumulate in model always, therefore term vector is not yet
It is applicable in.So we have proposed combine the iteration of word vector to expand convolutional neural networks+condition random field model.
Steps are as follows for specific method: the vector for obtaining each word first indicates.In next step, vector expression is input to expansion
In convolutional neural networks, after four layers of expansion convolutional layer, the feature and a vector for obtaining each word are indicated.It connects down
Come, the expression of these vectors is input in condition random field, condition random field exports its label.Details is as follows:
1. reading in data set CCKS2017 first.In data set, every style of writing originally includes two parts, word, corresponding label.Null
For dividing different training samples.After the sample of data set is upset at random, data set is divided by training with the ratio of 8:1
Collection and test set.
2. model, including three parts are constructed, word vector, expansion convolutional neural networks, condition random field.Word vector is used for
Distribution to each word indicates that expansion convolutional neural networks are used to extract the feature of each word, and condition random field is in training rank
Section is used for assessment tag subsequence score, is used to export the sequence label of highest scoring in test phase.Wherein, word vector module is adopted
With the term vector that pre-training is crossed on external corpus.
3., by word vector module, obtaining the expression of its vector, input using every 32 samples of training set as a batch
Into model, model is trained.In order to avoid over-fitting, we are added dropout layers after module, with certain probability
(being set as 0.5 herein) inactivates term vector.Trained objective function is to minimize negative log-likelihood function, it may be assumed that
Wherein:
Wherein s (x, y) is score function.
4. repeating big totally 100 epoch of step 3.After the completion of training, model parameter is saved in local file.It reads and surveys
Examination collection data, the prediction name entity on test set, and labeled data is compared, measure model performance.Test index uses F1 value,
It is defined as follows:
F1 value=(2* accurate rate * recall rate)/(accurate rate+recall rate)
The entity number that accurate rate=(the correctly predicted number in the entity that prediction obtains)/prediction obtains
Recall rate=(number being predicted correctly in the entity of labeled data)/labeled data entity number
In order to embody the effect of our models, we select other two model to compare.One is two-way length
Phase memory network+condition random field (BiLSTM+CRF), the model are a classical models for naming Entity recognition field, are being permitted
Good effect is all achieved on more data sets.Another model is HITSZ_CNER, which is in CCKS2017 match
Champion's model, the i.e. optimal model of achievement in match.
Test result is as shown in table 2, we compared the effect of our model (IDCNN+CRF) and previous model,
Generally, our model on electronic health record text name Entity recognition work on have biggish promotion, it is each not
It is generic physically, also have bigger promotion.In addition, we compared using radical feature (with feat) and not
Using the modelling effect of feature (no feat), comparison display, the radical feature that we are arranged can be obviously improved model performance.This
Invention is from the feature of case history text, the methods of reasonable utilization word vector, radical feature, expansion convolutional neural networks,
Preferably to be identified to medical bodies.
Specific structure of the invention is as shown in Fig. 2.
Described in attached drawing positional relationship for only for illustration, should not be understood as the limitation to this patent;
Table 2. compares the effect (F1 value, %) of different models.
Method | Bilstm+CRF | HITSZ_CNER | IDCNN+CRFw/ofeat. | IDCNN+CRF |
Organ | 88.10 | 87.42 | 87.10 | 87.56 |
Symptom | 95.73 | 96.34 | 95.16 | 96.94 |
Disease | 77.45 | 78.60 | 79.57 | 80.14 |
It checks | 95.69 | 94.36 | 96.02 | 96.11 |
Treatment method | 72.71 | 78.92 | 75.10 | 75.74 |
It amounts to | 90.82 | 91.08 | 91.62 | 92.53 |
The same or similar label correspond to the same or similar components;
Described in attached drawing positional relationship for only for illustration, should not be understood as the limitation to this patent;
Obviously, the above embodiment of the present invention be only to clearly illustrate example of the present invention, and not be pair
The restriction of embodiments of the present invention.For those of ordinary skill in the art, may be used also on the basis of the above description
To make other variations or changes in different ways.There is no necessity and possibility to exhaust all the enbodiments.It is all this
Made any modifications, equivalent replacements, and improvements etc., should be included in the claims in the present invention within the spirit and principle of invention
Protection scope within.
Claims (4)
1. a kind of case history text based on iteration expansion convolutional neural networks names entity recognition method, which is characterized in that including
Following steps:
S1: the model for naming the iteration of Entity recognition to expand convolutional neural networks and condition random field is established;
S2: the loss function of model is established;
S3: the training of model is carried out, and is tested on test set.
2. the case history text according to claim 1 based on iteration expansion convolutional neural networks names entity recognition method,
It is characterized in that, the detailed process of the step S1 is:
S11: building Embedding, since model will handle text, and text cannot directly be handled by model, and needing first will be literary
Word is converted into vector expression, i.e., with Embedding layers of completion, the vector including word is indicated and the vector of its radical indicates;
S12: building iteration expands convolutional neural networks, and for extracting feature, expansion convolutional neural networks include four layers of expansion volume
Lamination, expansion radius are respectively 1,2,3,3, and every layer includes 100 convolution kernels, and each convolution kernel width is 3, the last layer it is defeated
It is re-entered into first layer out, i.e. our so-called iteration, altogether iteration 4 times;
S13: building conditional random field models, the feature that previous step is extracted is as the input of condition random field, condition random
Field exports a sequence label to each word, to mark whether the word is a part of entity, if it does, being opening for entity
Head, intermediate or ending, and belong to what kind of entity.
3. the case history text according to claim 2 based on iteration expansion convolutional neural networks names entity recognition method,
It is characterized in that, the detailed process of the step S2 is:
S21: loss function is provided by negative log-likelihood function, and likelihood function is equal to, and the score of the label come is predicted, than upper
The score of all possible label:
Wherein, s (x, y) is score function;
S22: the calculating of score is divided into two parts, conversion score A i, j between (1) label are transformed into label j's from label i
Score;(2) label score Pm, the n of word, that is, give some word m, and label is the score of n, it may be assumed that
4. the case history text according to claim 3 based on iteration expansion convolutional neural networks names entity recognition method,
It is characterized in that, detailed process is as follows by the step S3:
S31: to input text, individual character processing is split, each word obtains its radical, by Embedding layers of acquisition word and radical
Vector indicate, be input to iteration expansion convolutional neural networks and to extract feature the feature extracted be input to condition random
In, final label is obtained;
S32: label and the model answer of prediction are compared, loss function is calculated by mode described in step S2, uses Adam
Optimizer, Lai Youhua loss function update model parameter;
S33: by data set according to ratio cut partition training set, the test set of 9:1, repetitive exercise 100 times on training set, at 50 times
Penalty values have tended towards stability afterwards, the model after training are saved, and the test result on test set.When test result, to test
Each word of sample all exports a corresponding label;
S34: S31-S33 is repeated, 5 cross validations are done on test set, are weighed using indexs such as accurate rate, recall rate, F1 values
The effect for measuring model, takes 5 average values as last effect.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811563980.0A CN109670179B (en) | 2018-12-20 | 2018-12-20 | Medical record text named entity identification method based on iterative expansion convolutional neural network |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811563980.0A CN109670179B (en) | 2018-12-20 | 2018-12-20 | Medical record text named entity identification method based on iterative expansion convolutional neural network |
Publications (2)
Publication Number | Publication Date |
---|---|
CN109670179A true CN109670179A (en) | 2019-04-23 |
CN109670179B CN109670179B (en) | 2022-11-11 |
Family
ID=66144168
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201811563980.0A Active CN109670179B (en) | 2018-12-20 | 2018-12-20 | Medical record text named entity identification method based on iterative expansion convolutional neural network |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109670179B (en) |
Cited By (18)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110321566A (en) * | 2019-07-10 | 2019-10-11 | 北京邮电大学 | Chinese name entity recognition method, device, computer equipment and storage medium |
CN110334186A (en) * | 2019-07-08 | 2019-10-15 | 北京三快在线科技有限公司 | Data query method, apparatus, computer equipment and computer readable storage medium |
CN110377903A (en) * | 2019-06-24 | 2019-10-25 | 浙江大学 | A kind of Sentence-level entity and relationship combine abstracting method |
CN110502742A (en) * | 2019-07-11 | 2019-11-26 | 中国科学院计算技术研究所 | A kind of complexity entity abstracting method, device, medium and system |
CN110674641A (en) * | 2019-10-06 | 2020-01-10 | 武汉鸿名科技有限公司 | GPT-2 model-based Chinese electronic medical record entity identification method |
CN110716991A (en) * | 2019-10-11 | 2020-01-21 | 掌阅科技股份有限公司 | Method for displaying entity associated information based on electronic book and electronic equipment |
CN110837736A (en) * | 2019-11-01 | 2020-02-25 | 浙江大学 | Character structure-based named entity recognition method for Chinese medical record of iterative expansion convolutional neural network-conditional random field |
CN111079377A (en) * | 2019-12-03 | 2020-04-28 | 哈尔滨工程大学 | Method for recognizing named entities oriented to Chinese medical texts |
CN111180026A (en) * | 2019-12-23 | 2020-05-19 | 卫宁健康科技集团股份有限公司 | Special diagnosis and treatment view system and method |
CN111222340A (en) * | 2020-01-15 | 2020-06-02 | 东华大学 | Breast electronic medical record entity recognition system based on multi-standard active learning |
CN111339764A (en) * | 2019-09-18 | 2020-06-26 | 华为技术有限公司 | Chinese named entity recognition method and device |
WO2020252950A1 (en) * | 2019-06-17 | 2020-12-24 | 五邑大学 | Named entity recognition method for medical texts based on pre-training model and fine turning technology |
CN112487807A (en) * | 2020-12-09 | 2021-03-12 | 重庆邮电大学 | Text relation extraction method based on expansion gate convolution neural network |
WO2021146831A1 (en) * | 2020-01-20 | 2021-07-29 | 京东方科技集团股份有限公司 | Entity recognition method and apparatus, dictionary creation method, device, and medium |
CN113204970A (en) * | 2021-06-07 | 2021-08-03 | 吉林大学 | BERT-BilSTM-CRF named entity detection model and device |
CN113836926A (en) * | 2021-09-27 | 2021-12-24 | 北京林业大学 | Electronic medical record named entity identification method, electronic equipment and storage medium |
CN113963304A (en) * | 2021-12-20 | 2022-01-21 | 山东建筑大学 | Cross-modal video time sequence action positioning method and system based on time sequence-space diagram |
CN116821286A (en) * | 2023-08-23 | 2023-09-29 | 北京宝隆泓瑞科技有限公司 | Correlation rule analysis method and system for gas pipeline accidents |
Citations (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106446526A (en) * | 2016-08-31 | 2017-02-22 | 北京千安哲信息技术有限公司 | Electronic medical record entity relation extraction method and apparatus |
CN107220237A (en) * | 2017-05-24 | 2017-09-29 | 南京大学 | A kind of method of business entity's Relation extraction based on convolutional neural networks |
US20170308790A1 (en) * | 2016-04-21 | 2017-10-26 | International Business Machines Corporation | Text classification by ranking with convolutional neural networks |
CN107315734A (en) * | 2017-05-04 | 2017-11-03 | 中国科学院信息工程研究所 | A kind of method and system for becoming pronouns, general term for nouns, numerals and measure words standardization based on time window and semanteme |
CN107527073A (en) * | 2017-09-05 | 2017-12-29 | 中南大学 | The recognition methods of entity is named in electronic health record |
CN107977361A (en) * | 2017-12-06 | 2018-05-01 | 哈尔滨工业大学深圳研究生院 | The Chinese clinical treatment entity recognition method represented based on deep semantic information |
CN108229582A (en) * | 2018-02-01 | 2018-06-29 | 浙江大学 | Entity recognition dual training method is named in a kind of multitask towards medical domain |
CN108563626A (en) * | 2018-01-22 | 2018-09-21 | 北京颐圣智能科技有限公司 | Medical text name entity recognition method and device |
CN108829681A (en) * | 2018-06-28 | 2018-11-16 | 北京神州泰岳软件股份有限公司 | A kind of name entity extraction method and device |
CN108920460A (en) * | 2018-06-26 | 2018-11-30 | 武大吉奥信息技术有限公司 | A kind of training method and device of the multitask deep learning model of polymorphic type Entity recognition |
-
2018
- 2018-12-20 CN CN201811563980.0A patent/CN109670179B/en active Active
Patent Citations (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20170308790A1 (en) * | 2016-04-21 | 2017-10-26 | International Business Machines Corporation | Text classification by ranking with convolutional neural networks |
CN106446526A (en) * | 2016-08-31 | 2017-02-22 | 北京千安哲信息技术有限公司 | Electronic medical record entity relation extraction method and apparatus |
CN107315734A (en) * | 2017-05-04 | 2017-11-03 | 中国科学院信息工程研究所 | A kind of method and system for becoming pronouns, general term for nouns, numerals and measure words standardization based on time window and semanteme |
CN107220237A (en) * | 2017-05-24 | 2017-09-29 | 南京大学 | A kind of method of business entity's Relation extraction based on convolutional neural networks |
CN107527073A (en) * | 2017-09-05 | 2017-12-29 | 中南大学 | The recognition methods of entity is named in electronic health record |
CN107977361A (en) * | 2017-12-06 | 2018-05-01 | 哈尔滨工业大学深圳研究生院 | The Chinese clinical treatment entity recognition method represented based on deep semantic information |
CN108563626A (en) * | 2018-01-22 | 2018-09-21 | 北京颐圣智能科技有限公司 | Medical text name entity recognition method and device |
CN108229582A (en) * | 2018-02-01 | 2018-06-29 | 浙江大学 | Entity recognition dual training method is named in a kind of multitask towards medical domain |
CN108920460A (en) * | 2018-06-26 | 2018-11-30 | 武大吉奥信息技术有限公司 | A kind of training method and device of the multitask deep learning model of polymorphic type Entity recognition |
CN108829681A (en) * | 2018-06-28 | 2018-11-16 | 北京神州泰岳软件股份有限公司 | A kind of name entity extraction method and device |
Non-Patent Citations (1)
Title |
---|
杨锦锋等: "电子病历命名实体识别和实体关系抽取研究综述", 《自动化学报》 * |
Cited By (27)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2020252950A1 (en) * | 2019-06-17 | 2020-12-24 | 五邑大学 | Named entity recognition method for medical texts based on pre-training model and fine turning technology |
CN110377903A (en) * | 2019-06-24 | 2019-10-25 | 浙江大学 | A kind of Sentence-level entity and relationship combine abstracting method |
CN110377903B (en) * | 2019-06-24 | 2020-08-14 | 浙江大学 | Sentence-level entity and relation combined extraction method |
CN110334186B (en) * | 2019-07-08 | 2021-09-28 | 北京三快在线科技有限公司 | Data query method and device, computer equipment and computer readable storage medium |
CN110334186A (en) * | 2019-07-08 | 2019-10-15 | 北京三快在线科技有限公司 | Data query method, apparatus, computer equipment and computer readable storage medium |
CN110321566B (en) * | 2019-07-10 | 2020-11-13 | 北京邮电大学 | Chinese named entity recognition method and device, computer equipment and storage medium |
CN110321566A (en) * | 2019-07-10 | 2019-10-11 | 北京邮电大学 | Chinese name entity recognition method, device, computer equipment and storage medium |
CN110502742A (en) * | 2019-07-11 | 2019-11-26 | 中国科学院计算技术研究所 | A kind of complexity entity abstracting method, device, medium and system |
CN111339764A (en) * | 2019-09-18 | 2020-06-26 | 华为技术有限公司 | Chinese named entity recognition method and device |
CN110674641B (en) * | 2019-10-06 | 2024-02-02 | 湖北大学 | Chinese electronic medical record entity identification method based on GPT-2 model |
CN110674641A (en) * | 2019-10-06 | 2020-01-10 | 武汉鸿名科技有限公司 | GPT-2 model-based Chinese electronic medical record entity identification method |
WO2021068932A1 (en) * | 2019-10-11 | 2021-04-15 | 掌阅科技股份有限公司 | Method based on electronic book for presenting information associated with entity |
CN110716991A (en) * | 2019-10-11 | 2020-01-21 | 掌阅科技股份有限公司 | Method for displaying entity associated information based on electronic book and electronic equipment |
CN110837736A (en) * | 2019-11-01 | 2020-02-25 | 浙江大学 | Character structure-based named entity recognition method for Chinese medical record of iterative expansion convolutional neural network-conditional random field |
CN110837736B (en) * | 2019-11-01 | 2021-08-10 | 浙江大学 | Named entity recognition method of Chinese medical record based on word structure |
CN111079377B (en) * | 2019-12-03 | 2022-12-13 | 哈尔滨工程大学 | Method for recognizing named entities of Chinese medical texts |
CN111079377A (en) * | 2019-12-03 | 2020-04-28 | 哈尔滨工程大学 | Method for recognizing named entities oriented to Chinese medical texts |
CN111180026A (en) * | 2019-12-23 | 2020-05-19 | 卫宁健康科技集团股份有限公司 | Special diagnosis and treatment view system and method |
CN111222340A (en) * | 2020-01-15 | 2020-06-02 | 东华大学 | Breast electronic medical record entity recognition system based on multi-standard active learning |
WO2021146831A1 (en) * | 2020-01-20 | 2021-07-29 | 京东方科技集团股份有限公司 | Entity recognition method and apparatus, dictionary creation method, device, and medium |
CN112487807A (en) * | 2020-12-09 | 2021-03-12 | 重庆邮电大学 | Text relation extraction method based on expansion gate convolution neural network |
CN112487807B (en) * | 2020-12-09 | 2023-07-28 | 重庆邮电大学 | Text relation extraction method based on expansion gate convolutional neural network |
CN113204970A (en) * | 2021-06-07 | 2021-08-03 | 吉林大学 | BERT-BilSTM-CRF named entity detection model and device |
CN113836926A (en) * | 2021-09-27 | 2021-12-24 | 北京林业大学 | Electronic medical record named entity identification method, electronic equipment and storage medium |
CN113963304B (en) * | 2021-12-20 | 2022-06-28 | 山东建筑大学 | Cross-modal video time sequence action positioning method and system based on time sequence-space diagram |
CN113963304A (en) * | 2021-12-20 | 2022-01-21 | 山东建筑大学 | Cross-modal video time sequence action positioning method and system based on time sequence-space diagram |
CN116821286A (en) * | 2023-08-23 | 2023-09-29 | 北京宝隆泓瑞科技有限公司 | Correlation rule analysis method and system for gas pipeline accidents |
Also Published As
Publication number | Publication date |
---|---|
CN109670179B (en) | 2022-11-11 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109670179A (en) | Case history text based on iteration expansion convolutional neural networks names entity recognition method | |
CN109920501B (en) | Electronic medical record classification method and system based on convolutional neural network and active learning | |
WO2021151353A1 (en) | Medical entity relationship extraction method and apparatus, and computer device and readable storage medium | |
Yu et al. | Automatic ICD code assignment of Chinese clinical notes based on multilayer attention BiRNN | |
CN110069779B (en) | Symptom entity identification method of medical text and related device | |
CN112597774B (en) | Chinese medical named entity recognition method, system, storage medium and equipment | |
CN108628824A (en) | A kind of entity recognition method based on Chinese electronic health record | |
CN108614885A (en) | Knowledge mapping analysis method based on medical information and device | |
CN110069631A (en) | A kind of text handling method, device and relevant device | |
CN110675944A (en) | Triage method and device, computer equipment and medium | |
CN109378066A (en) | A kind of control method and control device for realizing disease forecasting based on feature vector | |
CN116682553B (en) | Diagnosis recommendation system integrating knowledge and patient representation | |
CN110427486B (en) | Body condition text classification method, device and equipment | |
CN111048167A (en) | Hierarchical case structuring method and system | |
CN112489769A (en) | Intelligent traditional Chinese medicine diagnosis and medicine recommendation system for chronic diseases based on deep neural network | |
JP7464800B2 (en) | METHOD AND SYSTEM FOR RECOGNITION OF MEDICAL EVENTS UNDER SMALL SAMPLE WEAKLY LABELING CONDITIONS - Patent application | |
CN112420191A (en) | Traditional Chinese medicine auxiliary decision making system and method | |
CN110428907A (en) | A kind of text mining method and system based on unstructured electronic health record | |
CN116910172B (en) | Follow-up table generation method and system based on artificial intelligence | |
CN114781382A (en) | Medical named entity recognition system and method based on RWLSTM model fusion | |
Ke et al. | Medical entity recognition and knowledge map relationship analysis of Chinese EMRs based on improved BiLSTM-CRF | |
CN109360658A (en) | A kind of the disease pattern method for digging and device of word-based vector model | |
CN112071431B (en) | Clinical path automatic generation method and system based on deep learning and knowledge graph | |
CN111627561B (en) | Standard symptom extraction method, device, electronic equipment and storage medium | |
Chen et al. | Automatically structuring on Chinese ultrasound report of cerebrovascular diseases via natural language processing |
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 |