CN110110334A - A kind of remote medical consultation with specialists recording text error correction method based on natural language processing - Google Patents
A kind of remote medical consultation with specialists recording text error correction method based on natural language processing Download PDFInfo
- Publication number
- CN110110334A CN110110334A CN201910379327.7A CN201910379327A CN110110334A CN 110110334 A CN110110334 A CN 110110334A CN 201910379327 A CN201910379327 A CN 201910379327A CN 110110334 A CN110110334 A CN 110110334A
- Authority
- CN
- China
- Prior art keywords
- text
- word
- debugging
- error correction
- processed
- 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
- G06F40/232—Orthographic correction, e.g. spell checking or vowelisation
-
- 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
- G16H80/00—ICT specially adapted for facilitating communication between medical practitioners or patients, e.g. for collaborative diagnosis, therapy or health monitoring
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Computational Linguistics (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Biomedical Technology (AREA)
- Artificial Intelligence (AREA)
- Pathology (AREA)
- Epidemiology (AREA)
- Primary Health Care (AREA)
- Public Health (AREA)
- Machine Translation (AREA)
- Debugging And Monitoring (AREA)
Abstract
The invention discloses a kind of remote medical consultation with specialists recording text error correction method based on natural language processing, belong to big data technical field, including deployment central server and several clients, preprocessing module is established in central server, database, debugging module and correction module, solves the technical issues of text automatic errordetecting automatic error-correcting, the present invention dissipates string progress debugging using CRF and n-gram and carries out error correction according to specific error reason, very high level is reached in the ability of error correction, and it increases in accuracy than the prior art, the present invention can mitigate the pressure of work to the staff of remote diagnosis section, improve the efficiency of work.
Description
Technical field
The invention belongs to big data technical fields more particularly to a kind of remote medical consultation with specialists based on natural language processing to record text
This error correction method.
Background technique
With the development of computer technology and network technology, Telemedicine Consultation has become the one of modern medical service system system
A important component part.There is long-range section staff typing as consultation of doctors result is single, will appear multiword in Input Process, lacks
Word, misspelling, it is therefore desirable to there is special manpower or system to check and proofread these texts.At present to tele-medicine
The proof-reading of the consultation of doctors opinion list of record during the consultation of doctors is not only time-consuming but also laborious still based on artificial, so right
Telemedicine Consultation process carries out automatic Proofreading at the text information of middle formation to have great importance in telemedicine field.
Summary of the invention
The object of the present invention is to provide a kind of the remote medical consultation with specialists recording text error correction method based on natural language processing, solution
The technical issues of text automatic errordetecting automatic error-correcting.
To achieve the above object, the present invention adopts the following technical scheme:
A kind of remote medical consultation with specialists recording text error correction method based on natural language processing, includes the following steps:
Step 1: deployment central server and several clients, established in central server preprocessing module, database,
Debugging module and correction module, all clients pass through internet communication with central server;
Step 2: multiple urtext being inputted by any client, urtext is sent to center service by client
Device, central server stores all urtext into database, and is established in the database for storing and accumulating original
The tranining database of beginning text;
Step 3: the urtext in tranining database being classified as right-on text and Error Text, to completely just
True text and Error Text is segmented and is divided word processing, according to the errors present and type of error of corpus in urtext
Training corpus is marked, setting mark C is represented correctly, and mark R represents redundancy, and mark D represents missing, is marked O and is represented accidentally generation, mark
M represents missing;
CRF condition random field is called, and obtains training pattern using training corpus;
Step 4: text to be processed being inputted by any client, client is by File Transfer to be processed to center service
Device, the preprocessing module in central server pre-process text to be processed comprising following steps:
Step A1: segmented and divided text to be processed to word processing;
Step A2: by the participle in text to be processed and word is divided to be labeled as testing material;
Step 5: the debugging module in central server carries out debugging to text to be processed, and its step are as follows:
Step B1: debugging is carried out to the testing material in text to be processed according to training pattern and CRF condition random field, is obtained
To CRF condition random field debugging result;
Step B2: traversing all scattered strings in text to be processed, carries out n-gram to text to be processed and dissipates string debugging, obtains
N-gram dissipates string debugging result;
Step B3: fusion conditions random field debugging result and n-gram dissipate string debugging as a result, carrying out to text to be processed
Mark, obtains the final result of text debugging;
Step 6: the final result for the text debugging that step 5 obtains is input to correction module, error correction mould by central server
Block carries out error correction to the final result of text debugging, and its step are as follows:
Step C1: building language model corrects missing errors;
Step C2: the word marked containing redundancy error or word are directly deleted;
Step C3: it is corrected using word of the homonym dictionary to the label containing wrong generation in text, completes oneself of text
Dynamic error correction;
Step C4: output corrected text;
Step 7: establishing text proofreading main interface, text debugging interface and text error correction interface, center service in client
Device transmits the final result of text debugging, corrected text and text to be processed to client, and client is in text school
To main interface, text debugging interface and text error correction interface show respectively text to be processed, text debugging final result and entangle
Wrong text.
Preferably, when executing step 3, right-on text and Error Text are segmented using the library SnowNLP
With divide word processing.
Preferably, when executing step C1, language model selects three gram language models, and three gram language models are expressed as i-th
Word w on a positioniWith two word w of fronti-1And wi-2Related, formula indicates are as follows:
Wherein P indicates conditional probability;S indicates current statement or symbol string;N indicates that front and back character string number takes 3 here.
Preferably, the correction for being carried out missing errors to the final result of text debugging using three gram language models, is needed just
The corpus of true marking error type and errors present, its step are as follows:
Step S1: segmenting the final result of the text debugging of input, and finds out missing mark M;
Step S2: the previous and the latter word for marking M is extracted, and is recorded in missing text;
Step S3: traversing three gram language models of building, judge lack text recorded in word whether with dictionary sentence
In it is first identical with third word: if it is not the same, then error correction fails, and continues to search next missing label, repeatedly hold
Row step S1 to step S3, until missing text is all corrected;If identical, step S5 is executed;
Step S5: judge whether unique: if unique, second word of selected sentence is lacked word;Such as
Fruit is not unique, selects second word in the higher sentence of word frequency as lacked word;Error correction success, continues to search next
Missing label, repeats step S1 to step S5, until missing text is all corrected.
Preferably, it when executing step C3, is entangled using word of the homonym dictionary to the label containing wrong generation in text
Just, its step are as follows:
Step T1: text is segmented, and text after traversal participle is searched accidentally for error label O, marks the previous word of O
Belong to accidentally for mistake, which is carried out to the mark of phonetic, and record phonetic;
Step T2: judge to exist in constructed homonym dictionary with the presence or absence of the phonetic: if it does not, illustrating that error correction is lost
It loses;If it does, illustrating the mistake pronoun language, there are homonyms, using these homonyms as accidentally for candidate word;
Step T3: all homonyms accidentally for candidate word are successively substituted into prototype statement, calculate separately the probability of sentence simultaneously
Arrange in descending order, using sort first sentence in homonym as accidentally pronoun language correction word.
A kind of remote medical consultation with specialists recording text error correction method based on natural language processing of the present invention, solves text
The technical issues of automatic errordetecting automatic error-correcting, the present invention carry out debugging and according to the original that specifically malfunctions using the scattered string of CRF and n-gram
Because carrying out error correction, very high level is reached in the ability of error correction, and increase in accuracy than the prior art, this hair
The bright pressure that can mitigate work to the staff of remote diagnosis section, improves the efficiency of work.
Detailed description of the invention
Fig. 1 is general flow chart of the invention;
Fig. 2 is CRF debugging flow chart of the invention;
Fig. 3 is that n-gama of the invention dissipates string debugging flow chart;
Fig. 4 is the building flow chart of language model of the invention;
Fig. 5 is the building flow chart of homonym dictionary of the invention.
Specific embodiment
A kind of remote medical consultation with specialists recording text error correction method based on natural language processing as Figure 1-Figure 5, including such as
Lower step:
Step 1: deployment central server and several clients, established in central server preprocessing module, database,
Debugging module and correction module, all clients pass through internet communication with central server;
Step 2: multiple urtext being inputted by any client, urtext is sent to center service by client
Device, central server stores all urtext into database, and is established in the database for storing and accumulating original
The tranining database of beginning text;
Step 3: the urtext in tranining database being classified as right-on text and Error Text, to completely just
True text and Error Text is segmented and is divided word processing, according to the errors present and type of error of corpus in urtext
Training corpus is marked, setting mark C is represented correctly, and mark R represents redundancy, and mark D represents missing, is marked O and is represented accidentally generation, mark
M represents missing;
CRF condition random field is called, and obtains training pattern using training corpus;
Step 4: text to be processed being inputted by any client, client is by File Transfer to be processed to center service
Device, the preprocessing module in central server pre-process text to be processed comprising following steps:
Step A1: segmented and divided text to be processed to word processing;
Step A2: by the participle in text to be processed and word is divided to be labeled as testing material;
Step 5: the debugging module in central server carries out debugging to text to be processed, and its step are as follows:
Step B1: debugging is carried out to the testing material in text to be processed according to training pattern and CRF condition random field, is obtained
To CRF condition random field debugging result;
CRF condition random field is a kind of discriminate probabilistic model, using CRF condition random field by text as word sequence or
Word sequence is analyzed, with X={ x1x2...xnIndicate the sequence of observations, Y={ y1y2...ynIndicate flag sequence, it has ready conditions
The available characteristic function of random field:
Wherein i indicates the position where in observation sequence sentence, zxIndicate the standardization of the observation sequence marked, p table
Show that the probability of prediction error type, λ indicate the weight assigned;Which characteristic function j indicates;F indicates characteristic function;
Step B2: traversing all scattered strings in text to be processed, carries out n-gram to text to be processed and dissipates string debugging, obtains
N-gram dissipates string debugging result;
N-gram dissipate string debugging pass through natural language symbol string in n symbol and meanwhile probability of occurrence statistical data come
The structural relation for inferring sentence, uses wiThe linguistic notation currently to be occurred is indicated, in prediction wiProbability of occurrence when, need to consider
The linguistic notation of front n-1, is formulated as p (wi|wi-n+1wi-n+2wi-n+3...wi-1), n takes 3 herein.
Step B3: fusion conditions random field debugging result and n-gram dissipate string debugging as a result, carrying out to text to be processed
Mark, obtains the final result of text debugging;
Step 6: the final result for the text debugging that step 5 obtains is input to correction module, error correction mould by central server
Block carries out error correction to the final result of text debugging, and its step are as follows:
Step C1: building language model corrects missing errors;
Step C2: the word marked containing redundancy error or word are directly deleted;
Step C3: it is corrected using word of the homonym dictionary to the label containing wrong generation in text, completes oneself of text
Dynamic error correction;
Step C4: output corrected text;
Step 7: establishing text proofreading main interface, text debugging interface and text error correction interface, center service in client
Device transmits the final result of text debugging, corrected text and text to be processed to client, and client is in text school
To main interface, text debugging interface and text error correction interface show respectively text to be processed, text debugging final result and entangle
Wrong text.
Preferably, when executing step 3, right-on text and Error Text are segmented using the library SnowNLP
With divide word processing.
Preferably, when executing step C1, language model selects three gram language models, and three gram language models are expressed as i-th
Word w on a positioniWith two word w of fronti-1And wi-2Related, formula indicates are as follows:
Wherein P indicates conditional probability;S indicates current statement or symbol string;N indicates that front and back character string number takes 3 here.
Preferably, the correction for being carried out missing errors to the final result of text debugging using three gram language models, is needed just
The corpus of true marking error type and errors present, its step are as follows:
Step S1: segmenting the final result of the text debugging of input, and finds out missing mark M;
Step S2: the previous and the latter word for marking M is extracted, and is recorded in missing text;
Step S3: traversing three gram language models of building, judge lack text recorded in word whether with dictionary sentence
In it is first identical with third word: if it is not the same, then error correction fails, and continues to search next missing label, repeatedly hold
Row step S1 to step S3, until missing text is all corrected;If identical, step S5 is executed;
Step S5: judge whether unique: if unique, second word of selected sentence is lacked word;Such as
Fruit is not unique, selects second word in the higher sentence of word frequency as lacked word;Error correction success, continues to search next
Missing label, repeats step S1 to step S5, until missing text is all corrected.
Preferably, it when executing step C3, is entangled using word of the homonym dictionary to the label containing wrong generation in text
Just, its step are as follows:
Step T1: text is segmented, and text after traversal participle is searched accidentally for error label O, marks the previous word of O
Belong to accidentally for mistake, which is carried out to the mark of phonetic, and record phonetic;
Step T2: judge to exist in constructed homonym dictionary with the presence or absence of the phonetic: if it does not, illustrating that error correction is lost
It loses;If it does, illustrating the mistake pronoun language, there are homonyms, using these homonyms as accidentally for candidate word;
Step T3: all homonyms accidentally for candidate word are successively substituted into prototype statement, calculate separately the probability of sentence simultaneously
Arrange in descending order, using sort first sentence in homonym as accidentally pronoun language correction word.
A kind of remote medical consultation with specialists recording text error correction method based on natural language processing of the present invention, solves text
The technical issues of automatic errordetecting automatic error-correcting, the present invention carry out debugging and according to the original that specifically malfunctions using the scattered string of CRF and n-gram
Because carrying out error correction, very high level is reached in the ability of error correction, and increase in accuracy than the prior art, this hair
The bright pressure that can mitigate work to the staff of remote diagnosis section, improves the efficiency of work.
Claims (5)
1. a kind of remote medical consultation with specialists recording text error correction method based on natural language processing, characterized by the following steps:
Step 1: deployment central server and several clients establish preprocessing module, database, debugging in central server
Module and correction module, all clients pass through internet communication with central server;
Step 2: multiple urtext are inputted by any client, urtext is sent to central server by client,
Central server stores all urtext into database, and is established in the database for storing and accumulating original text
This tranining database;
Step 3: the urtext in tranining database being classified as right-on text and Error Text, to right-on
Text and Error Text are segmented and are divided word processing, are marked according to the errors present of corpus in urtext and type of error
Training corpus, setting mark C are represented correctly, and mark R represents redundancy, and mark D represents missing, are marked O and are represented accidentally generation, mark M generation
Table missing;
CRF condition random field is called, and obtains training pattern using training corpus;
Step 4: text to be processed is inputted by any client, client by File Transfer to be processed to central server, in
Preprocessing module in central server pre-processes text to be processed comprising following steps:
Step A1: segmented and divided text to be processed to word processing;
Step A2: by the participle in text to be processed and word is divided to be labeled as testing material;
Step 5: the debugging module in central server carries out debugging to text to be processed, and its step are as follows:
Step B1: debugging is carried out to the testing material in text to be processed according to training pattern and CRF condition random field, is obtained
CRF condition random field debugging result;
Step B2: traversing all scattered strings in text to be processed, carries out n-gram to text to be processed and dissipates string debugging, obtains n-
Gram dissipates string debugging result;
Step B3: fusion conditions random field debugging result and n-gram dissipate string debugging as a result, being labeled to text to be processed,
Obtain the final result of text debugging;
Step 6: the final result for the text debugging that step 5 obtains is input to correction module, correction module pair by central server
The final result of text debugging carries out error correction, and its step are as follows:
Step C1: building language model corrects missing errors;
Step C2: the word marked containing redundancy error or word are directly deleted;
Step C3: it is corrected using word of the homonym dictionary to the label containing wrong generation in text, completes entangling automatically for text
Wrong function;
Step C4: output corrected text;
Step 7: establishing text proofreading main interface, text debugging interface and text error correction interface in client, central server will
Final result, corrected text and the text to be processed of text debugging are transmitted to client, and client is in text proofreading master
Interface, text debugging interface and text error correction interface show text to be processed, the final result of text debugging and error correction text respectively
This.
2. a kind of remote medical consultation with specialists recording text error correction method based on natural language processing as described in claim 1, feature
It is: when executing step 3, is segmented and divided word processing to right-on text and Error Text using the library SnowNLP.
3. a kind of remote medical consultation with specialists recording text error correction method based on natural language processing as described in claim 1, feature
Be: when executing step C1, language model selects three gram language models, and three gram language models are expressed as on i-th of position
Word wiWith two word w of fronti-1And wi-2Related, formula indicates are as follows:
Wherein P indicates conditional probability;S indicates while statement or symbol string;The number of n expression symbol.
4. a kind of remote medical consultation with specialists recording text error correction method based on natural language processing as claimed in claim 3, feature
It is: carries out the correction of missing errors to the final result of text debugging using three gram language models, need correct marking error
The corpus of type and errors present, its step are as follows:
Step S1: segmenting the final result of the text debugging of input, and finds out missing mark M;
Step S2: the previous and the latter word for marking M is extracted, and is recorded in missing text;
Step S3: traversing three gram language models of building, whether judges to lack word recorded in text in dictionary sentence the
One is identical with third word: if it is not the same, then error correction fails, and continues to search next missing label, repeating step
Rapid S1 to step S3, until missing text is all corrected;If identical, step S5 is executed;
Step S5: judge whether unique: if unique, second word of selected sentence is lacked word;If no
Uniquely, select second word in the higher sentence of word frequency as lacked word;Error correction success, continues to search next missing
Label, repeats step S1 to step S5, until missing text is all corrected.
5. a kind of remote medical consultation with specialists recording text error correction method based on natural language processing as described in claim 1, feature
It is: when executing step C3, is corrected using word of the homonym dictionary to the label containing wrong generation in text, step is such as
Under:
Step T1: text is segmented, and text after traversal participle is searched accidentally for error label O, the previous word for marking O belongs to
Accidentally for mistake, which is carried out to the mark of phonetic, and records phonetic;
Step T2: judge to exist in constructed homonym dictionary with the presence or absence of the phonetic: if it does not, illustrating that error correction fails;
If it does, illustrating the mistake pronoun language, there are homonyms, using these homonyms as accidentally for candidate word;
Step T3: all homonyms accidentally for candidate word are successively substituted into prototype statement, calculate separately the probability of sentence and by drop
Sequence arrangement, using sort first sentence in homonym as mistake pronoun language correction word.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910379327.7A CN110110334B (en) | 2019-05-08 | 2019-05-08 | Remote consultation record text error correction method based on natural language processing |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910379327.7A CN110110334B (en) | 2019-05-08 | 2019-05-08 | Remote consultation record text error correction method based on natural language processing |
Publications (2)
Publication Number | Publication Date |
---|---|
CN110110334A true CN110110334A (en) | 2019-08-09 |
CN110110334B CN110110334B (en) | 2022-09-13 |
Family
ID=67488670
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910379327.7A Active CN110110334B (en) | 2019-05-08 | 2019-05-08 | Remote consultation record text error correction method based on natural language processing |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110110334B (en) |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110532562A (en) * | 2019-08-30 | 2019-12-03 | 联想(北京)有限公司 | Neural network training method, Chinese idiom misuse detection method, device and electronic equipment |
CN110765836A (en) * | 2019-08-28 | 2020-02-07 | 云知声智能科技股份有限公司 | Text positioning method and system based on natural language understanding |
CN110826304A (en) * | 2019-11-13 | 2020-02-21 | 北京雅丁信息技术有限公司 | Medical corpus labeling method |
CN113627191A (en) * | 2021-07-05 | 2021-11-09 | 中国气象局公共气象服务中心(国家预警信息发布中心) | Automatic labeling method and system for meteorological early warning sample semantics |
CN116975298A (en) * | 2023-09-22 | 2023-10-31 | 厦门智慧思明数据有限公司 | NLP-based modernized society governance scheduling system and method |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105279149A (en) * | 2015-10-21 | 2016-01-27 | 上海应用技术学院 | Chinese text automatic correction method |
CN107122346A (en) * | 2016-12-28 | 2017-09-01 | 平安科技(深圳)有限公司 | The error correction method and device of a kind of read statement |
-
2019
- 2019-05-08 CN CN201910379327.7A patent/CN110110334B/en active Active
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105279149A (en) * | 2015-10-21 | 2016-01-27 | 上海应用技术学院 | Chinese text automatic correction method |
CN107122346A (en) * | 2016-12-28 | 2017-09-01 | 平安科技(深圳)有限公司 | The error correction method and device of a kind of read statement |
WO2018120889A1 (en) * | 2016-12-28 | 2018-07-05 | 平安科技(深圳)有限公司 | Input sentence error correction method and device, electronic device, and medium |
Non-Patent Citations (1)
Title |
---|
韩彦昭等: "基于条件随机场模型和文本纠错的微博新词词性识别研究", 《南京大学学报(自然科学)》 * |
Cited By (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110765836A (en) * | 2019-08-28 | 2020-02-07 | 云知声智能科技股份有限公司 | Text positioning method and system based on natural language understanding |
CN110765836B (en) * | 2019-08-28 | 2022-04-29 | 云知声智能科技股份有限公司 | Text positioning method and system based on natural language understanding |
CN110532562A (en) * | 2019-08-30 | 2019-12-03 | 联想(北京)有限公司 | Neural network training method, Chinese idiom misuse detection method, device and electronic equipment |
CN110532562B (en) * | 2019-08-30 | 2021-07-16 | 联想(北京)有限公司 | Neural network training method, idiom misuse detection method and device and electronic equipment |
CN110826304A (en) * | 2019-11-13 | 2020-02-21 | 北京雅丁信息技术有限公司 | Medical corpus labeling method |
CN113627191A (en) * | 2021-07-05 | 2021-11-09 | 中国气象局公共气象服务中心(国家预警信息发布中心) | Automatic labeling method and system for meteorological early warning sample semantics |
CN116975298A (en) * | 2023-09-22 | 2023-10-31 | 厦门智慧思明数据有限公司 | NLP-based modernized society governance scheduling system and method |
CN116975298B (en) * | 2023-09-22 | 2023-12-05 | 厦门智慧思明数据有限公司 | NLP-based modernized society governance scheduling system and method |
Also Published As
Publication number | Publication date |
---|---|
CN110110334B (en) | 2022-09-13 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110110334A (en) | A kind of remote medical consultation with specialists recording text error correction method based on natural language processing | |
CN106202153B (en) | A kind of the spelling error correction method and system of ES search engine | |
CN100511215C (en) | Multilingual translation memory and translation method thereof | |
CN111353306B (en) | Entity relationship and dependency Tree-LSTM-based combined event extraction method | |
CN109145260B (en) | Automatic text information extraction method | |
CN109471793B (en) | Webpage automatic test defect positioning method based on deep learning | |
US9110852B1 (en) | Methods and systems for extracting information from text | |
CN107526721B (en) | Ambiguity elimination method and device for comment vocabularies of e-commerce products | |
CN104317882A (en) | Decision-based Chinese word segmentation and fusion method | |
CN110598787B (en) | Software bug classification method based on self-defined step length learning | |
CN114970502B (en) | Text error correction method applied to digital government | |
Langenecker et al. | Towards learned metadata extraction for data lakes | |
CN115935914A (en) | Admission record missing text supplementing method | |
JP5812534B2 (en) | Question answering apparatus, method, and program | |
CN106484676A (en) | Biological Text protein reference resolution method based on syntax tree and domain features | |
CN112818693A (en) | Automatic extraction method and system for electronic component model words | |
CN113190692A (en) | Self-adaptive retrieval method, system and device for knowledge graph | |
CN110909174B (en) | Knowledge graph-based method for improving entity link in simple question answering | |
CN117422074A (en) | Method, device, equipment and medium for standardizing clinical information text | |
US9146918B2 (en) | Compressing data for natural language processing | |
CN114021572B (en) | Natural language processing method, device, equipment and readable storage medium | |
CN113468311B (en) | Knowledge graph-based complex question and answer method, device and storage medium | |
CN109657207B (en) | Formatting processing method and processing device for clauses | |
CN114238595A (en) | Metallurgical knowledge question-answering method and system based on knowledge graph | |
CN112685434A (en) | Operation and maintenance question-answering method based on knowledge graph |
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 |