CN111209728B - Automatic labeling and inputting method for test questions - Google Patents
Automatic labeling and inputting method for test questions Download PDFInfo
- Publication number
- CN111209728B CN111209728B CN202010032609.2A CN202010032609A CN111209728B CN 111209728 B CN111209728 B CN 111209728B CN 202010032609 A CN202010032609 A CN 202010032609A CN 111209728 B CN111209728 B CN 111209728B
- Authority
- CN
- China
- Prior art keywords
- keywords
- question
- test
- test question
- keyword
- 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
- 238000012360 testing method Methods 0.000 title claims abstract description 200
- 238000002372 labelling Methods 0.000 title claims abstract description 17
- 238000000034 method Methods 0.000 title claims abstract description 14
- 230000011218 segmentation Effects 0.000 claims abstract description 12
- 238000010276 construction Methods 0.000 claims description 26
- 230000014509 gene expression Effects 0.000 claims description 9
- 238000012216 screening Methods 0.000 claims description 7
- 238000013179 statistical model Methods 0.000 abstract description 5
- 238000000605 extraction Methods 0.000 description 3
- 238000001914 filtration Methods 0.000 description 3
- 238000012545 processing Methods 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 238000012549 training Methods 0.000 description 1
Abstract
The invention relates to an automatic labeling and inputting method for test questions, which comprises the following steps: automatically converting a word test question document into a corresponding html document and storing the html document; automatically extracting text content of an html document; performing word segmentation on the text content by adopting an n-gram model in the statistical model, and performing part-of-speech tagging on the segmented words by using a hidden Markov model; extracting test question features of words marked with parts of speech, judging whether the features of the words correspond to the features of preset test question keywords, if so, marking the words with keywords and marking the positions of the words in the text; dividing the test questions one by one according to the question number keywords; and assembling the test questions in a test question constructing module according to the keywords segmented by questions and the positions of the keywords in the text, and storing the assembled test questions to the corresponding positions. The invention effectively improves the accuracy of automatically inputting test questions and solves the problem that pictures or formulas cannot be automatically input.
Description
Technical Field
The invention relates to the field of online examination, in particular to an automatic labeling and inputting method for test questions.
Background
With the development of network technology, more and more education and training institutions adopt an online examination mode, so that the establishment of an examination question library is a core suitable for the scale, the scientificity and the standardization of the examination, and the first essential link of the establishment of the examination question library is the input of examination questions.
The existing test question inputting method includes inputting test questions manually and inputting based on regular expressions, wherein the manual input usually includes a plurality of text boxes in a specific test question inputting system webpage, corresponding contents such as a test question stem, an answer, options, analysis and the like are required to be respectively input, if the test questions comprise pictures and formulas, the input is also required to be performed by clicking a mode of inserting the pictures or formulas, however, the mode of inputting the test questions manually is complicated, and when the number of the test questions is large, a great amount of time is required to be consumed for inputting, and batch input of the test questions cannot be realized. The method is mainly applied to the test questions of a pure text type, and the method is not applicable to the extraction of test questions containing pictures or input of a pure text formula.
Disclosure of Invention
The invention aims to overcome at least one defect in the prior art, and provides an automatic marking and inputting method for test questions, which can realize batch inputting of the test questions and improve the accuracy of inputting the test questions.
The technical scheme adopted by the invention is as follows:
the automatic labeling and inputting method for the test questions comprises the following steps:
s1, receiving a word test question document, automatically converting the word test question document into a corresponding html document, and storing the html document;
s2, automatically extracting text content of the html document;
s3, carrying out word segmentation on the extracted text content, carrying out part-of-speech tagging on the segmented words, and tagging the positions of the words in the text;
s4, extracting test question features of the words marked with the parts of speech, judging whether the features of the words correspond to the features of preset test question keywords, if so, marking the words, and taking the words marked as the keywords as determined keywords;
s5, dividing the test questions one by one according to the determined keywords, and storing the keywords of each test question;
s6, assembling the test questions in a test question construction module according to the keywords segmented by questions and the positions of the keywords in the text, and storing the assembled test questions to the corresponding positions according to the keywords and the positions of the keywords in the text.
By converting word documents into html documents for processing, pictures and formulas in test questions can be completely saved, and the problem that the test question documents with the pictures or formulas cannot be automatically input is solved; the statistical model is adopted to segment the words of the test question text, the word parts of the segmented words are labeled, and the word parts are used as auxiliary judgment conditions to extract the characteristic keywords of the test question, so that the accuracy of automatic test question segmentation is improved.
Further, in the step S1, the step of automatically converting the word test question document into a corresponding html document also includes converting a formula in the document into a picture, recording a corresponding position tag of the picture in the document, and storing the picture and the corresponding position tag thereof and the html document in the same folder.
Further, the step S2 further includes extracting a position tag of the picture and marking a corresponding position of the picture in the text.
Further, in the step S4, whether the feature of the word corresponds to the feature of the preset test question keyword is specifically that the multi-layer regular expression is adopted to extract the test question feature of the word, whether the word corresponds to the preset test question keyword feature is judged, if yes, keyword marking is carried out on the word, and the word marked as the keyword is used as the determined keyword.
And extracting test question features of the words through the multi-layer regular expressions, judging whether the words correspond to preset test question keyword features or not, and if so, marking the words with keywords, so that the test question features can be well judged, and the accuracy of marking the test question keywords is improved.
Further, the step S4 further includes performing a second filtering on the words marked as keywords.
Further, the determined keywords include one or more of a question number, an option, an answer, and a resolution.
Further, in the step S6, the question-by-question segmentation is performed on the test questions according to the determined keywords, specifically, the question number keywords are taken as judgment basis, and if answer keywords and/or analysis keywords are included between the two question number keywords, the test questions are segmented; if no answer key and/or analysis key exists between one question key and the next question key, marking the next question key as interference.
And further, performing secondary screening on the marked keywords, namely combining all keywords in a single test question through a random variation algorithm, calculating the scores of all keyword combinations, and selecting the keyword with the highest score as a determined keyword.
Combining and secondarily screening the marked keywords through a random variation algorithm, calculating the score of the keyword combination, and selecting the keyword with the highest score as the determined keyword, so that the most suitable combination can be found, and the accuracy of extracting the test question features is improved.
Further, the step S6 further includes assembling the determined keywords in the test question construction modules of different types, calculating the scores of the assembled test questions in the test question construction modules of different types, comparing the scores of the same test question in the test question construction modules of different types, taking the type of the test question construction module with the highest score as the type of the test question, and marking the type of the test question.
The keyword with the highest score is assembled in the test question construction modules of different types, the scores of the same test question in the test question construction modules of different types are compared, the type of the test question construction module with the highest score is selected as the type of the test question, and the accuracy of automatic marking of the test question type is improved.
Further, the step S6 further includes searching for a position tag of the picture in the test question, and storing the picture included in the test question to a corresponding position.
Compared with the prior art, the invention has the beneficial effects that:
(1) By converting the word document into the html document for processing, the pictures and formulas in the test questions can be completely saved, the corresponding positions of the pictures and formulas in the text are marked, when the test questions are called, the pictures are found through indexes and the corresponding positions are inserted, and the automatic input of the test question document with the pictures and formulas is realized.
(2) The method comprises the steps of segmenting the test question content by adopting an n-gram model in a statistical model, and marking the parts of speech of the segmented words by adopting a hidden Markov model. The part of speech is used as auxiliary judgment condition to extract the characteristic key words of the test questions, thereby improving the accuracy of automatic segmentation of the test questions.
(3) The words are subjected to test question feature extraction by adopting the multi-layer regular expression, and the limiting conditions are wider based on the word parts as the basis for judging the test question keywords, so that each feature in the test questions is well reserved, and the test question feature extraction is more flexible compared with that of a single regular expression.
(4) By adopting the random variation algorithm to divide the single test question key word characteristics, combining and secondary screening, selecting the key word combination with the highest score, the accuracy of extracting the test question characteristics is improved, assembling the test questions in the test question construction modules of different types by using the combination with the highest score of each test question key word, selecting the type of the test question construction module with the highest score as the type of the test question, and improving the accuracy of automatic labeling of the test question type.
Drawings
FIG. 1 is an overall flow chart of an embodiment of the present invention;
FIG. 2 is a keyword labeling judgment flow chart according to an embodiment of the present invention, wherein a is a question number keyword labeling judgment flow chart, b is an option keyword labeling judgment flow chart, c is an answer keyword labeling judgment flow chart, and d is an analysis keyword labeling judgment flow chart;
FIG. 3 is a flow chart of test question assembly according to an embodiment of the present invention.
Detailed Description
The drawings are for illustrative purposes only and are not to be construed as limiting the invention. For better illustration of the following embodiments, some parts of the drawings may be omitted, enlarged or reduced, and do not represent the actual product dimensions; it will be appreciated by those skilled in the art that certain well-known structures in the drawings and descriptions thereof may be omitted.
As shown in fig. 1, this embodiment provides a method for automatically labeling and inputting test questions, including:
s1, receiving a word test question document, automatically converting the word test question document into a corresponding html document, and storing the html document;
s2, automatically extracting text content of the html document;
s3, segmenting the words in the extracted text content, marking the parts of speech of the segmented words, and marking the positions of the words in the text;
s4, extracting test question features of the words marked with the parts of speech, judging whether the features of the words correspond to the features of preset test question keywords, if so, marking the words, and taking the words marked as the keywords as determined keywords;
s5, dividing the test questions one by one according to the determined keywords, and storing the keywords of each test question;
s6, assembling the test questions in a test question construction module according to the keywords segmented by questions and the positions of the keywords in the text, finding out corresponding documents according to the keywords and the positions of the keywords in the text, and storing the documents in the corresponding positions.
Specifically, when a test question is required to be recorded, the system reads the content of a word document uploaded by a user through a program, converts the word document into an html document, acquires the text content of the html document by using an html analyzer, then performs word segmentation on the text content by adopting an n-gram model in a statistical model, marks the corresponding position of the segmented word in the text and generates a word position index, performs part-of-speech labeling on the segmented word by adopting a hidden Markov model, and places the word with the labeled part-of-speech into a word set; sequentially transmitting the marked words into a test question feature keyword marking module to extract test question features of the words marked with word parts, comparing the features of the words with the features of test question keywords according to preset conditions, judging whether the words are test question keywords or not, wherein the specific judging process is to sequentially judge whether the conditions are met according to the preset conditions from top to bottom, if one condition is not met, the current words are not the corresponding features of the test question keywords, if the words meeting all the conditions are met, the words are marked with keywords, the positions of the words in texts are marked with keywords to generate keyword indexes, and the words marked with the keywords are used as determined keywords; and finally, dividing the test questions one by one according to the determined question number keywords, storing the keywords of each test question, assembling each test question in a test question construction module according to the keywords divided one by one and the positions of the keywords in the texts, finding out the corresponding texts through the keywords and the position indexes of the keywords contained in the keywords, and storing the texts in the appointed positions. By converting the word document into the html document for processing in the implementation manner, the pictures and formulas in the test questions can be completely saved, and the problem that the test question document with the pictures or formulas cannot be automatically input is solved; the statistical model is adopted to segment the words of the test question text, the word parts of the segmented words are labeled, and the word parts are used as auxiliary judgment conditions to extract the characteristic keywords of the test question, so that the accuracy of automatic test question segmentation is improved.
In this embodiment, in the step S1, the step of automatically converting the word test question document into the corresponding html document further includes converting a formula in the document into a picture, recording a corresponding position tag of the picture in the document, and storing the picture and the corresponding position tag thereof and the html document in the same folder.
In this embodiment, the step S2 further includes extracting a position tag of the picture and marking a corresponding position of the picture in the text.
Specifically, in the step S1, the automatic conversion of the word test question document into the corresponding html document also includes converting the formula in the document into the picture, recording the corresponding position tag of the picture in the html document, and then placing all the pictures and the corresponding position tags thereof in the same folder stored with the html document, thereby being capable of recording the test question document with the picture and the formula in the label without losing the content information.
In this embodiment, in the step S4, whether the feature of the word corresponds to the feature of the preset test question keyword is specifically that a multi-layer regular expression is adopted to extract the test question feature of the word, whether the word corresponds to the preset test question keyword feature is judged, if yes, keyword marking is performed on the word, and the word marked as the keyword is used as the determined keyword.
And extracting test question features of the words through the multi-layer regular expressions, judging whether the words correspond to preset test question keyword features or not, and if so, marking the words with keywords, so that the test question features can be well judged, and the accuracy of marking the test question keywords is improved.
In this embodiment, the step S4 further includes performing a second filtering on the labeled keyword.
In this embodiment, the determined keywords include one or more of question number, option, answer, and resolution.
Specifically, the determined keywords include one or more of question numbers, options, answers and analysis, in the step S4, whether the features of the words are corresponding to the features of the preset test question keywords is specifically that the labeled words are sequentially transmitted into a test question feature keyword labeling module, the test question feature keyword labeling module extracts test question features of the words by adopting a multi-layer regular expression, the features of the words are compared with the features of the test question keywords according to preset conditions, whether the words are test question keywords is judged, the preset keyword condition judgment flow chart is specifically shown as fig. 2, wherein an a chart is a judgment flow chart of the question number keyword labels, a b chart is a judgment flow chart of the option keyword labels, a c chart is a judgment flow chart of the answer keyword labels, a d chart is a judgment flow chart of the keyword labels, the keyword judgment process is specifically that whether the words with labeled word parts sequentially from top to bottom meet the keyword judgment conditions, if all the words are met, the words with the first condition is labeled, the words with the first condition is not met, the corresponding words with the second condition is not met, and the second condition is still labeled, and therefore the feature is still needed to be labeled, and the judgment is carried out by the second condition.
In this embodiment, in the step S6, the question-by-question segmentation is performed on the test questions according to the determined keywords, specifically, the question number keywords are taken as the judgment basis, and if answer keywords and/or analysis keywords are included between the two question number keywords, the test questions are segmented; if no answer key and/or analysis key exists between one question key and the next question key, marking the next question key as interference.
Specifically, in the step S6, the question-by-question segmentation is performed on the test questions according to the determined keywords, specifically, the question-number keywords are taken as judgment basis, answer keywords or analysis keywords are needed to be included between the two question-number keywords to segment the test questions, when no answer keywords or analysis keywords exist between one question-number keyword and the next question-number keyword, the second question-number keyword is marked as interference, the interference keywords are ignored to continue indexing to find the next question-number keyword, until the question-number keywords meeting the conditions are found, and after the question-by-question segmentation is completed, the keywords of each test question and the corresponding keyword position indexes are stored; through the implementation mode, the situation that the test questions are divided into errors due to the fact that the keywords are marked in errors can be avoided.
In this embodiment, the second filtering of the labeled keywords is specifically to combine all keywords in a single test question with keywords by a random variation algorithm and calculate scores of all keyword combinations.
Specifically, a single test question contains a plurality of characteristic keywords, but the keywords obtained through the previous labeling are not necessarily correct, and secondary screening is needed for the characteristic keywords of the single test question, wherein the secondary screening adopts the steps of combining all the keywords in the single test question through a random variation algorithm, then calculating the combined scores of all the keywords, and selecting the combined keyword with the highest score as a determined keyword.
Combining and secondarily screening the marked keywords through a random variation algorithm, calculating the score of the keyword combination, and selecting the keyword with the highest score as the determined keyword, so that the most suitable combination can be found, and the accuracy of extracting the test question features is improved.
In this embodiment, the step S6 further includes selecting a keyword combination with the highest score to perform test question assembly in different types of test question construction modules, calculating scores of the assembled test questions in different types of test question construction modules, comparing scores of the same test question in different types of test question construction modules, taking a type of the test question construction module with the highest score as a type of the test question, and marking the type of the test question.
Specifically, the types of the test question constructing modules comprise a single-choice question constructing module, a multi-choice question constructing module, a judging question constructing module, a simple answer constructing module and an analyzing question constructing module, the keyword combination with the highest score is respectively put into the test question constructing modules of different types for test question assembly, as shown in fig. 3, a flow chart of the test question constructing module for assembling test questions is shown, the scores of the same test question in the test question constructing modules of different types are calculated, and specifically, in the single-choice question constructing module, the more conditions that the unique question number, the unique answer, the unique option sequence appear and the analytic answer are met, the higher the calculated scores are; in the multi-choice question constructing module, the more the conditions that the question number is unique, the answers are multiple, no characters exist among the answers, the options appear in sequence and the only analysis is satisfied, the higher the calculated score is; in the judging question constructing module, the more conditions of unique question numbers, unique answers and unique analysis are satisfied, the higher the calculated score is; in the simple answer construction module, the more conditions that the question number is unique, the answer is unique and the analysis is unique are satisfied, the higher the calculated score is; in the analysis question construction module, the more the condition that the number of questions is multiple, the answer is unique and the analysis is unique is satisfied, the higher the calculated score is. And finally, finding out a corresponding text through the keywords and the position indexes of the keywords contained in the keywords and storing the corresponding text to the corresponding position, storing the assembled test questions and the test question types in a one-to-one correspondence manner, searching the corresponding position of the corresponding picture mark in the test question, correspondingly storing the picture contained in the test question, and facilitating calling and inputting.
The keyword with the highest score is combined in the test question construction modules of different types to carry out test question assembly and calculate the score, the scores of the same test question in the test question construction modules of different types are compared, the type of the test question construction module with the highest score is selected as the type of the test question, and the accuracy of automatic marking of the test question type is improved.
In this embodiment, step S6 further includes searching for a position tag of a picture in the test question, and storing the picture included in the test question to a corresponding position.
Specifically, after the types of the test questions are determined, the types of the test questions and the test questions are stored in a one-to-one correspondence mode, position labels of corresponding pictures in the test questions are searched, and the pictures contained in the test questions are stored in the corresponding positions in the test questions, so that the input is conveniently invoked and performed.
Claims (1)
1. The automatic labeling and inputting method for the test questions is characterized by comprising the following steps of:
s1, receiving a word test question document, automatically converting the word test question document into a corresponding html document, and storing the html document;
s2, automatically extracting text content of the html document;
s3, carrying out word segmentation on the extracted text content, carrying out part-of-speech tagging on the segmented words, and tagging the positions of the words in the text;
s4, extracting test question features of the words marked with the parts of speech, judging whether the features of the words correspond to the features of preset test question keywords, if so, marking the words, and taking the words marked as the keywords as determined keywords;
s5, dividing the test questions one by one according to the determined keywords, and storing the keywords of each test question;
s6, assembling the test questions in a test question construction module according to the keywords segmented by questions and the positions of the keywords in the text, and storing the assembled test questions to corresponding positions according to the keywords and the positions of the keywords in the text;
the step S6 further comprises the steps of assembling the determined keywords in test question construction modules of different types, calculating the scores of the assembled test questions in the test question construction modules of different types, comparing the scores of the same test question in the test question construction modules of different types, taking the type of the test question construction module with the highest score as the type of the test question, and marking the type of the test question;
in the step S1, the word test question document is automatically converted into a corresponding html document, meanwhile, a formula in the document is converted into a picture, a corresponding position label of the picture in the document is recorded, and the picture and the corresponding position label thereof and the html document are stored in the same folder;
the step S2 further comprises the steps of extracting a position label of the picture and marking the corresponding position of the picture in a text;
in the step S4, whether the characteristics of the word correspond to the characteristics of the preset test question keywords is specifically determined by extracting the test question characteristics of the word by adopting a multi-layer regular expression, whether the word corresponds to the preset test question keyword characteristics is determined, if yes, the word is labeled, and the word labeled as the keyword is used as the determined keyword;
the step S4 further comprises the step of carrying out secondary screening on the words marked as the keywords;
the determined keywords comprise one or more of question numbers, options, answers and parsing;
in the step S5, the question-by-question segmentation is performed on the test questions according to the determined keywords, specifically, the question number keywords are taken as judgment basis, and if answer keywords and/or analysis keywords are included between the two question number keywords, the test questions are segmented; if no answer keyword and/or analysis keyword exists between one question keyword and the next question keyword, marking the next question keyword as interference;
the words marked as keywords are secondarily screened, namely, all keywords in a single test question are subjected to keyword combination through a random variation algorithm, the scores of all keyword combinations are calculated, and the keyword with the highest score is selected as a determined keyword;
step S6 further comprises searching a position label of the picture in the test question, and storing the picture contained in the test question to a corresponding position.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010032609.2A CN111209728B (en) | 2020-01-13 | 2020-01-13 | Automatic labeling and inputting method for test questions |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010032609.2A CN111209728B (en) | 2020-01-13 | 2020-01-13 | Automatic labeling and inputting method for test questions |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111209728A CN111209728A (en) | 2020-05-29 |
| CN111209728B true CN111209728B (en) | 2024-01-30 |
Family
ID=70786103
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010032609.2A Active CN111209728B (en) | 2020-01-13 | 2020-01-13 | Automatic labeling and inputting method for test questions |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111209728B (en) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111832282B (en) * | 2020-07-16 | 2023-04-14 | 平安科技（深圳）有限公司 | External knowledge fused BERT model fine adjustment method and device and computer equipment |
| CN111737949B (en) * | 2020-07-22 | 2021-07-06 | 江西风向标教育科技有限公司 | Topic content extraction method and device, readable storage medium and computer equipment |
| CN111680669A (en) * | 2020-08-12 | 2020-09-18 | 江西风向标教育科技有限公司 | Test question segmentation method and system and readable storage medium |
| CN112052646B (en) * | 2020-08-27 | 2024-03-29 | 安徽聚戎科技信息咨询有限公司 | Text data labeling method |
| CN113065316A (en) * | 2021-03-26 | 2021-07-02 | 洛阳圣昂通网络科技有限公司 | Method for dynamically converting formal thumbnail file into html (hypertext markup language) and inputting question bank, selecting questions from question bank and composing draft and generating thumbnail file |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104199871A (en) * | 2014-08-19 | 2014-12-10 | 南京富士通南大软件技术有限公司 | High-speed test question inputting method for intelligent teaching |
| CN106802937A (en) * | 2016-12-30 | 2017-06-06 | 江苏中育优教科技发展有限公司 | The conversion method and system of Word document |
| CN109614594A (en) * | 2018-11-27 | 2019-04-12 | 浙江万朋教育科技股份有限公司 | A method of topic document is resolved into exam pool data |
-
2020
- 2020-01-13 CN CN202010032609.2A patent/CN111209728B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104199871A (en) * | 2014-08-19 | 2014-12-10 | 南京富士通南大软件技术有限公司 | High-speed test question inputting method for intelligent teaching |
| CN106802937A (en) * | 2016-12-30 | 2017-06-06 | 江苏中育优教科技发展有限公司 | The conversion method and system of Word document |
| CN109614594A (en) * | 2018-11-27 | 2019-04-12 | 浙江万朋教育科技股份有限公司 | A method of topic document is resolved into exam pool data |
Non-Patent Citations (1)
| Title |
|---|
| 俞婷 ; 符云清 ; .基于词法分析和XML技术的多媒体试题批量导入研究.计算机应用与软件.2016,第33卷(第06期),第134-137页. * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111209728A (en) | 2020-05-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111209728B (en) | Automatic labeling and inputting method for test questions | |
| CN109947836B (en) | English test paper structuring method and device | |
| CN110119786B (en) | Text topic classification method and device | |
| CN110795543B (en) | Unstructured data extraction method, device and storage medium based on deep learning | |
| CN107766371B (en) | Text information classification method and device | |
| CN100511215C (en) | Multilingual translation memory and translation method thereof | |
| CN112269901A (en) | Fault distinguishing and reasoning method based on knowledge graph | |
| CN111274239B (en) | Test paper structuring processing method, device and equipment | |
| CN108334493B (en) | Question knowledge point automatic extraction method based on neural network | |
| CN112434691A (en) | HS code matching and displaying method and system based on intelligent analysis and identification and storage medium | |
| CN109949637B (en) | Automatic answering method and device for objective questions | |
| CN110110334B (en) | Remote consultation record text error correction method based on natural language processing | |
| JP2007087397A (en) | Morphological analysis program, correction program, morphological analyzer, correcting device, morphological analysis method, and correcting method | |
| CN114218379B (en) | Attribution method for question answering incapacity of intelligent question answering system | |
| CN110276069A (en) | A kind of Chinese braille mistake automatic testing method, system and storage medium | |
| CN112115301A (en) | Video annotation method and system based on classroom notes | |
| CN115034218A (en) | Chinese grammar error diagnosis method based on multi-stage training and editing level voting | |
| CN115292461A (en) | Man-machine interaction learning method and system based on voice recognition | |
| CN111680669A (en) | Test question segmentation method and system and readable storage medium | |
| KR101794547B1 (en) | System and Method for Automatically generating of personal wordlist and learning-training word | |
| CN110765107B (en) | Question type identification method and system based on digital coding | |
| CN110362691B (en) | Syntax tree bank construction system | |
| WO2002075583A1 (en) | A learning automatic data extraction system | |
| CN114492460A (en) | Event causal relationship extraction method based on derivative prompt learning | |
| Farasyn et al. | Challenges in tagging and parsing spoken dialects of Dutch |
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 |