CN109086313A - One kind carrying out examination question based on inverse text similarity and orders rearrangement processed - Google Patents
One kind carrying out examination question based on inverse text similarity and orders rearrangement processed Download PDFInfo
- Publication number
- CN109086313A CN109086313A CN201810681320.6A CN201810681320A CN109086313A CN 109086313 A CN109086313 A CN 109086313A CN 201810681320 A CN201810681320 A CN 201810681320A CN 109086313 A CN109086313 A CN 109086313A
- Authority
- CN
- China
- Prior art keywords
- examination question
- topic
- word
- frequency
- similarity
- 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.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/30—Semantic analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
Examination question is carried out based on inverse text similarity the invention discloses one kind and orders rearrangement processed, include the following steps, it needs to carry out duplicate checking processing to every problem of upload when the topic of oneself is uploaded to original problem database system by proposition staff first, duplicate checking processing, which is taken, carry out in the way of element category to original exam pool, then the similarity of more every topic one by one is gone in corresponding original exam pool by element, when the proposition of oneself is uploaded to original exam pool simultaneously by multiple proposition staff, the examination question for first uploading to interim problem database system needs first to be compared with the examination question in interim problem database system, then examination question of the topic of interim exam pool together with formal problem database system compares duplicate checking.The invention designs advantages of simple, easy to operate, not only reduces manual labor, and improve examination question re-scheduling efficiency, while similar examination question can effectively be avoided to upload, and improves and examines topic quality, safety and stability is applied widely, is conducive to popularize.
Description
Technical field
The invention belongs to examination questions to order technical field processed, more specifically, more particularly to it is a kind of based on inverse text similarity into
Row examination question orders rearrangement processed.
Background technique
During examination question life is processed, the similarity of topic directly affects the quality of examination question, how to reduce the similarity of topic,
By the investigation and judgement of pure artificial mechanism formula, inefficiency is time-consuming and laborious, and the knowledge and ability of effect topic personnel on trial
Horizontal influence.To improve re-scheduling quality, examination question re-scheduling is carried out using computer, due to the original examination question of existing problem database system
Enormous amount calculates if the examination question uploaded needs to compare duplicate checking one by one and workload is very big, algorithm to re-scheduling and
The performance requirement of re-scheduling machine is very high, and traditional solution belongs to mechanical, artificial examination question and examines topic method, re-scheduling
The influence of the knowledge and ability level of effect topic personnel on trial, although the similarity analysis method energy based on character string pattern matching
The similar caused examination question replication problem of text is enough solved, but helpless for the examination question of semantic similarity, is unfavorable for extensively
Promotion and popularization.
Summary of the invention
The purpose of the present invention is to solve disadvantages existing in the prior art, and the one kind proposed is similar based on inverse text
Degree carries out examination question and orders rearrangement processed.
To achieve the above object, the invention provides the following technical scheme:
One kind carrying out examination question based on inverse text similarity and orders rearrangement processed, includes the following steps:
The per pass to upload is needed when S1, first proposition staff upload to the topic of oneself in original problem database system
Topic carries out duplicate checking processing, prevents from uploading same or similar topic type；
S2, duplicate checking processing are taken and carry out original exam pool by element category, in the way of the corresponding domain classification of topic, so
It need to only be determined when re-scheduling compares afterwards and compare element corresponding to topic, then go in corresponding original exam pool to compare one by one by element
The similarity of every topic；
S3, the similarity for judging two examination questions are first to construct high quality according to self-built test item bank and disclosed test item bank
Available corpus and participle library, and based on dictionary using segmenter to examination question carry out morphological analysis；
S4, the word frequency (TF) for then calculating examination question are calculated " inverse document frequency " (IDF) based on corpus, and are calculated with this
The TF-IDF of word；
S5, the TF-IDF again based on participle and word construct the feature vector of examination question, by the meaning of a word of examination question and the semanteme amount of progress
Change；
S6, finally by the feature vector of the examination question compared two-by-two using the cosine law calculate similarity, in this, as examination question
The foundation of similarity re-scheduling；
When the proposition of oneself is carried out uploading to original exam pool by S7, multiple proposition staff simultaneously, life in a period of time is taken
The proposition of topic personnel is all to upload to the mode of interim problem database system, and the examination question newly uploaded needs first with the examination in interim problem database system
Topic is compared, the examination question of the two propositions personnel for uploading to interim exam pool simultaneously, if there is similar or mutually sympathize with
Condition, is subject to first uploader；
Examination question of the topic of S8, then interim exam pool together with formal problem database system compares duplicate checking, is then incorporated into
In formal problem database system.
Preferably, the multiple tracks examination question uploaded in the step S1 for proposition staff, oneself needs to carry out at duplicate checking first
Reason, the topic for avoiding oneself from uploading have similar or same type topic.
Preferably, the calculating step in the step S4 is 1) to calculate word frequency: word frequency=appearance of some word in examination question
Number/article total degree；2) inverse document frequency (IDF) is calculated: inverse document frequency=log corpus examination question sum/comprising being somebody's turn to do
The examination question number+1 of word；3) term frequency-inverse document frequency (TF-IDF) is calculated: term frequency-inverse document frequency=word frequency * inverse document frequency.
Preferably, text is indicated in the step S5 in the form of numerical characteristics vector, mainly passes through two steps
To realize；1. creating a unique label for each word on entire examination question collection (containing many examination questions)；2. being every
A examination question constructs a feature vector, mainly includes frequency of occurrence of each word on document.
Preferably, four steps (1) are divided into the step S6 and count what all words in two comparison texts occurred respectively
Frequency, to obtain the corresponding vector of two texts；(2) the included angle cosine value of the two vectors is calculated using the cosine law；(3)
The range of the included angle cosine value of two vectors be 0-1,1 represent it is identical, 0 represent it is completely not identical；According to the threshold from setting
Value judges whether two texts are similar；(4) threshold value is set as 90%, and the examination question for regarding as comparing two-by-two if it is greater than 90% is
It is identical, the similarity value of examination question is returned if lower than 90%, is referred to for examination question uploader and careful topic personnel, is participated in be artificial
Judgement provides foundation.
It preferably, can be in the base of test item bank element in the step S8 in the optimization of the order of magnitude problem of calculating
The sub- element category that element is used on plinth, further decreases the number compared, accelerates the efficiency of re-scheduling.
Technical effect and advantage of the invention: provided by the invention a kind of based on inverse text similarity progress examination question life system row
Weighing method, compared with traditional method, the present invention first constructs available corpus and the participle library of high quality, and is based on dictionary
Morphological analysis is carried out to examination question using segmenter；The word frequency (TF) for calculating examination question calculates " inverse document frequency " based on corpus
(IDF), the TF-IDF of word and with this is calculated；The feature vector of TF-IDF building examination question based on participle and word, by the word of examination question
Justice and semanteme are quantified；The feature vector of the examination question compared two-by-two is calculated into similarity using the cosine law, in this, as examination
Inscribe the foundation of similarity re-scheduling；Based on the similarity analysis of " inverse text similarity ", can constantly be trained constantly more based on one
New corpus, similar caused examination question replication problem semantic to the text phase Sihe of examination question, can provide more accurate
More scientific similarity foundation, and interim exam pool can also be first passed through when uploading proposition, multiple proposition persons are uploaded
Examination question carry out preparatory re-scheduling, the examination question after re-scheduling is finally uploaded to formal exam pool, reduce repetitive rate that examination question uploads or
Similarity improves the efficiency of examination question re-scheduling by the setting of test item bank key element and sub- element, and invention design is simple to close
Reason, it is easy to operate, manual labor is not only reduced, and improve the efficiency of examination question re-scheduling, while can effectively avoid
Similar examination question is uploaded, improves and examines topic quality, safety and stability is applied widely, is conducive to popularize.
Specific embodiment
In order to make the objectives, technical solutions, and advantages of the present invention clearer, the present invention is carried out further detailed
Explanation.It should be appreciated that the specific embodiments described herein are only used to explain the present invention, it is not intended to limit the present invention.
Based on the specific embodiment in the present invention, those of ordinary skill in the art are obtained without making creative work
Every other embodiment, shall fall within the protection scope of the present invention.
One kind carrying out examination question based on inverse text similarity and orders rearrangement processed, includes the following steps:
The per pass to upload is needed when S1, first proposition staff upload to the topic of oneself in original problem database system
Topic carries out duplicate checking processing, prevents from uploading same or similar topic type；
S2, duplicate checking processing are taken and carry out original exam pool by element category, in the way of the corresponding domain classification of topic, so
It need to only be determined when re-scheduling compares afterwards and compare element corresponding to topic, then go in corresponding original exam pool to compare one by one by element
The similarity of every topic；
S3, the similarity for judging two examination questions are first to construct high quality according to self-built test item bank and disclosed test item bank
Available corpus and participle library, and based on dictionary using segmenter to examination question carry out morphological analysis；
S4, the word frequency (TF) for then calculating examination question are calculated " inverse document frequency " (IDF) based on corpus, and are calculated with this
The TF-IDF of word；
S5, the TF-IDF again based on participle and word construct the feature vector of examination question, by the meaning of a word of examination question and the semanteme amount of progress
Change；
S6, finally by the feature vector of the examination question compared two-by-two using the cosine law calculate similarity, in this, as examination question
The foundation of similarity re-scheduling；
When the proposition of oneself is carried out uploading to original exam pool by S7, multiple proposition staff simultaneously, life in a period of time is taken
The proposition of topic personnel is all to upload to the mode of interim problem database system, and the examination question newly uploaded needs first in interim problem database system
Examination question is compared, the examination question of the two propositions personnel for uploading to interim exam pool simultaneously, if there is similar or identical
Situation, is subject to first uploader；
Examination question of the topic of S8, then interim exam pool together with formal problem database system compares duplicate checking, is then incorporated into
In formal problem database system.
Further, the multiple tracks examination question uploaded in the step S1 for proposition staff, oneself needs to carry out duplicate checking first
Processing, the topic for avoiding oneself from uploading have similar or same type topic.
Further, the calculating step in the step S4 is 1) to calculate word frequency: word frequency=some word going out in examination question
Occurrence number/article total degree；2) calculating inverse document frequency (IDF): inverse document frequency ,=the examination question sum of log corpus/includes
The examination question number+1 of the word；3) calculate term frequency-inverse document frequency (TF-IDF): term frequency-inverse document frequency=word frequency * against document frequency
Rate.
Further, text is indicated in the form of numerical characteristics vector in the step S5, it is main to pass through two steps
It is rapid to realize；1. creating a unique label for each word on entire examination question collection (containing many examination questions)；2. being
Each examination question constructs a feature vector, mainly includes frequency of occurrence of each word on document.
Further, four steps (1) are divided into the step S6 and count all words appearance in two comparison texts respectively
Frequency, to obtain the corresponding vector of two texts；(2) the included angle cosine value of the two vectors is calculated using the cosine law；
The range of the included angle cosine value of (3) two vectors be 0-1,1 represent it is identical, 0 represent it is completely not identical；According to setting certainly
Whether two texts of threshold decision are similar；(4) threshold value is set as 90%, regards as the examination question compared two-by-two if it is greater than 90%
Be it is identical, if lower than the similarity value for returning to examination question if 90%, referred to for examination question uploader and careful topic personnel, for artificial ginseng
Foundation is provided with judgement.
It further, can be in test item bank element in the step S8 in the optimization of the order of magnitude problem of calculating
On the basis of use element sub- element category, further decrease the number compared, accelerate the efficiency of re-scheduling.
In summary: it is provided by the invention a kind of based on inverse text similarity progress examination question life rearrangement processed, with tradition
Method compare, the present invention first construct high quality available corpus and participle library, and based on dictionary use segmenter pair
Examination question carries out morphological analysis；The word frequency (TF) for calculating examination question calculates " inverse document frequency " (IDF) based on corpus, and in terms of this
Calculate the TF-IDF of word；The feature vector of TF-IDF building examination question based on participle and word, by the meaning of a word of examination question and the semanteme amount of progress
Change；The feature vector of the examination question compared two-by-two is calculated into similarity using the cosine law, in this, as examination question similarity re-scheduling
Foundation；Based on the similarity analysis of " inverse text similarity ", the corpus constantly updated can be constantly trained based on one, to examination
The semantic similar caused examination question replication problem of the text phase Sihe of topic, can provide more accurate more scientific similarity
Foundation, and can also first pass through interim exam pool when proposition uploading the examination question that multiple proposition persons upload is arranged in advance
Examination question after re-scheduling is finally uploaded to formal exam pool by weight, is reduced repetitive rate or similarity that examination question uploads, is passed through test item bank
The setting of key element and sub- element improves the efficiency of examination question re-scheduling, which designs advantages of simple, easy to operate, not only drops
Low manual labor, and the efficiency of examination question re-scheduling is improved, while can effectively avoid uploading similar examination question, it improves and examines
Quality is inscribed, safety and stability is applied widely, is conducive to popularize.
Finally, it should be noted that the foregoing is only a preferred embodiment of the present invention, it is not intended to restrict the invention,
Although the present invention is described in detail referring to the foregoing embodiments, for those skilled in the art, still may be used
To modify the technical solutions described in the foregoing embodiments or equivalent replacement of some of the technical features,
All within the spirits and principles of the present invention, any modification, equivalent replacement, improvement and so on should be included in of the invention
Within protection scope.
Claims (6)
1. one kind carries out examination question based on inverse text similarity and orders rearrangement processed, which comprises the steps of:
S1, first proposition staff need to carry out every problem of upload when the topic of oneself is uploaded to original problem database system
Duplicate checking processing, prevents from uploading same or similar topic type；
S2, duplicate checking processing are taken and carry out in the way of the corresponding domain classification of topic, so by element category to original exam pool
It need to only be determined when re-scheduling compares afterwards and compare element corresponding to topic, then go in corresponding original exam pool to compare one by one by element
The similarity of every topic；
S3, the similarity for judging two examination questions be first construct high quality according to self-built test item bank and disclosed test item bank can
Corpus and participle library, and morphological analysis is carried out to examination question using segmenter based on dictionary；
S4, the word frequency (TF) for then calculating examination question calculate " inverse document frequency " (IDF) based on corpus, and calculate word with this
TF-IDF；
S5, the TF-IDF again based on participle and word construct the feature vector of examination question, and the meaning of a word of examination question and semanteme are quantified；
S6, finally by the feature vector of the examination question compared two-by-two using the cosine law calculate similarity, it is similar in this, as examination question
Spend the foundation of re-scheduling；
When the proposition of oneself is uploaded to original exam pool simultaneously by S7, multiple proposition staff, proposition staff in a period of time is taken
Proposition is all to upload to the mode of interim problem database system, and the examination question newly uploaded needs first to carry out with the examination question in interim problem database system
It compares, the examination question of the two propositions personnel for uploading to interim exam pool simultaneously, if there is similar or same case, with elder generation
Subject to uploader；
Examination question of the topic of S8, then interim exam pool together with formal problem database system compares duplicate checking, is then incorporated into formal
Problem database system in.
2. according to claim 1 a kind of based on inverse text similarity progress examination question life rearrangement processed, it is characterised in that:
The multiple tracks examination question uploaded in the step S1 for proposition staff, oneself needs to carry out duplicate checking processing first, avoids oneself uploading
Topic have similar or same type topic.
3. according to claim 1 a kind of based on inverse text similarity progress examination question life rearrangement processed, it is characterised in that:
Calculating step in the step S4 is 1) to calculate word frequency: frequency of occurrence/article total degree of word frequency=some word in examination question；
2) inverse document frequency (IDF) is calculated: the inverse document frequency=log corpus examination question sum/examination question number+1 comprising the word；3) it counts
It calculates term frequency-inverse document frequency (TF-IDF): term frequency-inverse document frequency=word frequency * inverse document frequency.
4. according to claim 1 a kind of based on inverse text similarity progress examination question life rearrangement processed, it is characterised in that:
Text is indicated in the form of numerical characteristics vector in the step S5, is mainly realized by two steps；1. being entire
Each word on examination question collection (containing many examination questions) creates a unique label；2. constructing a spy for each examination question
Vector is levied, mainly includes frequency of occurrence of each word on document.
5. according to claim 1 a kind of based on inverse text similarity progress examination question life rearrangement processed, it is characterised in that:
It is divided into four steps (1) in the step S6 and counts the frequency that all words occur in two comparison texts respectively, to obtains two
The corresponding vector of a text；(2) the included angle cosine value of the two vectors is calculated using the cosine law；The angle of (3) two vectors
The range of cosine value be 0-1,1 represent it is identical, 0 represent it is completely not identical；According to two texts of threshold decision from setting
It is whether similar；(4) threshold value is set as 90%, the examination question for regarding as comparing two-by-two if it is greater than 90% be it is identical, if low
The similarity value that examination question is returned in 90% is referred to for examination question uploader and careful topic personnel, participates in judging to provide foundation to be artificial.
6. according to claim 1 a kind of based on inverse text similarity progress examination question life rearrangement processed, it is characterised in that:
For the son of element can be used on the basis of test item bank element in the optimization of the order of magnitude problem of calculating in the step S8
Element category further decreases the number compared, accelerates the efficiency of re-scheduling.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810681320.6A CN109086313A (en) | 2018-06-27 | 2018-06-27 | One kind carrying out examination question based on inverse text similarity and orders rearrangement processed |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810681320.6A CN109086313A (en) | 2018-06-27 | 2018-06-27 | One kind carrying out examination question based on inverse text similarity and orders rearrangement processed |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109086313A true CN109086313A (en) | 2018-12-25 |
Family
ID=64839917
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810681320.6A Pending CN109086313A (en) | 2018-06-27 | 2018-06-27 | One kind carrying out examination question based on inverse text similarity and orders rearrangement processed |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109086313A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110516232A (en) * | 2019-07-22 | 2019-11-29 | 北京师范大学 | A kind of automatic proposition method and system for Chinese evaluation and test |
| CN110990389A (en) * | 2019-11-29 | 2020-04-10 | 上海易点时空网络有限公司 | Method and device for simplifying question bank and computer readable storage medium |
| CN113051886A (en) * | 2021-03-25 | 2021-06-29 | 科大讯飞股份有限公司 | Test question duplicate checking method and device, storage medium and equipment |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105373594A (en) * | 2015-10-23 | 2016-03-02 | 广东小天才科技有限公司 | Method and apparatus for screening repeated test questions from question bank |
| CN105630767A (en) * | 2015-12-22 | 2016-06-01 | 北京奇虎科技有限公司 | Text similarity comparison method and device |
| CN105824798A (en) * | 2016-03-03 | 2016-08-03 | 云南电网有限责任公司教育培训评价中心 | Examination question de-duplicating method of examination question base based on examination question key word likeness |
| CN107832406A (en) * | 2017-11-03 | 2018-03-23 | 北京锐安科技有限公司 | Duplicate removal storage method, device, equipment and the storage medium of massive logs data |
-
2018
- 2018-06-27 CN CN201810681320.6A patent/CN109086313A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105373594A (en) * | 2015-10-23 | 2016-03-02 | 广东小天才科技有限公司 | Method and apparatus for screening repeated test questions from question bank |
| CN105630767A (en) * | 2015-12-22 | 2016-06-01 | 北京奇虎科技有限公司 | Text similarity comparison method and device |
| CN105824798A (en) * | 2016-03-03 | 2016-08-03 | 云南电网有限责任公司教育培训评价中心 | Examination question de-duplicating method of examination question base based on examination question key word likeness |
| CN107832406A (en) * | 2017-11-03 | 2018-03-23 | 北京锐安科技有限公司 | Duplicate removal storage method, device, equipment and the storage medium of massive logs data |
Non-Patent Citations (1)
| Title |
|---|
| YABEA: "TF-IDF算法（1）-算法概述", 《百度，HTTPS://WWW.CNBLOGS.COM/YBJOURNEY/P/4793370.HTML》 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110516232A (en) * | 2019-07-22 | 2019-11-29 | 北京师范大学 | A kind of automatic proposition method and system for Chinese evaluation and test |
| CN110516232B (en) * | 2019-07-22 | 2021-06-22 | 北京师范大学 | Automatic proposition method and system for Chinese evaluation |
| CN110990389A (en) * | 2019-11-29 | 2020-04-10 | 上海易点时空网络有限公司 | Method and device for simplifying question bank and computer readable storage medium |
| CN113051886A (en) * | 2021-03-25 | 2021-06-29 | 科大讯飞股份有限公司 | Test question duplicate checking method and device, storage medium and equipment |
| CN113051886B (en) * | 2021-03-25 | 2023-12-01 | 科大讯飞股份有限公司 | Test question duplicate checking method, device, storage medium and equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Hou et al. | Identification of tasks, datasets, evaluation metrics, and numeric scores for scientific leaderboards construction | |
| CN108573045B (en) | Comparison matrix similarity retrieval method based on multi-order fingerprints | |
| CN103631859B (en) | Intelligent review expert recommending method for science and technology projects | |
| CN109086313A (en) | One kind carrying out examination question based on inverse text similarity and orders rearrangement processed | |
| CN106844665B (en) | Thesis recommendation method based on reference relation distributed expression | |
| CN110674296B (en) | Information abstract extraction method and system based on key words | |
| Ding et al. | Research on text information mining technology of substation inspection based on improved Jieba | |
| Ansari et al. | Identifying semantically duplicate questions using data science approach: A quora case study | |
| Chi et al. | Application of BP neural network based on genetic algorithms optimization in prediction of postgraduate entrance examination | |
| Sah et al. | Machine Learning Approaches to Assess Mood of the News Editorial | |
| Sun et al. | Reca: Related tables enhanced column semantic type annotation framework | |
| Lan et al. | PCRED: Zero-shot relation triplet extraction with potential candidate relation selection and entity boundary detection | |
| Kumar et al. | NLP@ NISER: Classification of COVID19 tweets containing symptoms | |
| Zundi et al. | Word-level morpheme segmentation using transformer neural network | |
| Rao et al. | Keyword extraction in scientific documents | |
| Zhou et al. | Topic-Attentive Encoder-Decoder with Pre-Trained Language Model for Keyphrase Generation | |
| Sboev et al. | Evaluation of Machine Learning Methods for Relation Extraction Between Drug Adverse Effects and Medications in Russian Texts of Internet User Reviews | |
| Phuc et al. | Using SOM based graph clustering for extracting main ideas from documents | |
| Jo | Long text segmentation by string vector based KNN | |
| Wang et al. | Research on Semantic Consistency Problem Based on Word2vec | |
| Zhao et al. | Open domain event attribute extraction method | |
| Wang et al. | Attribute-based Sentiment Analysis for Recommendation | |
| Huang et al. | An automatic approach for extracting Chinese entity synonyms from encyclopedias | |
| Ning | Research on the extraction of accounting multi-relationship information based on cloud computing and multimedia | |
| Zhiming et al. | A topic detection method based on bicharacteristic vectors |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20181225 |