CN113312909B - Intelligent analysis test question answer method and system based on natural language processing - Google Patents
Intelligent analysis test question answer method and system based on natural language processing Download PDFInfo
- Publication number
- CN113312909B CN113312909B CN202110545942.8A CN202110545942A CN113312909B CN 113312909 B CN113312909 B CN 113312909B CN 202110545942 A CN202110545942 A CN 202110545942A CN 113312909 B CN113312909 B CN 113312909B
- Authority
- CN
- China
- Prior art keywords
- answers
- student
- answer
- sentences
- standard
- 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
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
- G06F40/289—Phrasal analysis, e.g. finite state techniques or chunking
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/205—Parsing
- G06F40/211—Syntactic parsing, e.g. based on context-free grammar [CFG] or unification grammars
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
Abstract
The invention discloses an intelligent test question analysis answer method and system based on natural language processing, which comprises the following steps: acquiring text data of test questions, including standard answers and student answers, and performing formalized processing; acquiring vector representation of each sentence based on a dynamic representation method of word vectors; based on the existing neural network model processed by natural language, a hybrid neural network model is established by utilizing an integrated thought; analyzing and comparing the obtained standard answers and the vector representations of the student answer sentences one by using a hybrid neural network model, and calculating the similarity of the sentences; and sequencing according to the similarity of the sentences, matching the student answers with the standard answers, analyzing the student answers according to the matching results, and giving corresponding answer guidance. The invention can help learners to know the insufficient answer, realize the teaching according to the material and help learners to improve the score.
Description
Technical Field
The invention relates to the technical field of natural language processing, in particular to an intelligent test question analysis answer method and system based on natural language processing.
Background
With the development of deep learning technology, the artificial intelligence technology is more mature. Natural language processing is part of artificial intelligence, and many excellent algorithm models are emerging, including neural network models such as CNN, RNN, transformer, GNN, etc., which can do more and more things and have better and more effects on natural language processing.
In the traditional analysis of test questions and answers, teachers are mainly used for guiding students, and individuation is lacking, namely, the teachers often neglect individual differences of learners and are difficult to teach according to the situation, so that the improvement of the performance of learners is not facilitated, and targeted answering guidance and suggestions cannot be obtained.
Disclosure of Invention
The first purpose of the invention is to overcome the defects and shortcomings of the prior art, and provide an intelligent analysis test question answer method based on natural language processing, which is used for carrying out similarity analysis on standard answers and student answers, matching the standard answers and the student answers, and finally giving corresponding answer guidance and suggestions according to matching results, so that students can know the defects of answering the questions.
The second objective of the present invention is to provide an intelligent analysis answer system for test questions based on natural language processing.
The first purpose of the invention is realized by the following technical scheme: an intelligent test question analysis and answer method based on natural language processing comprises the following steps:
s1, acquiring text data of test questions, including standard answers and student answers, and performing formalized processing on the text data;
s2, mapping words of the text data into vectors with unified dimensionality for representation based on a word vector dynamic representation method, and substituting the vectors into sentences to obtain vector representation of each sentence;
s3, establishing a hybrid neural network model by utilizing an integrated thought based on a neural network model processed by a natural language;
s4, based on the obtained vector representation of the text data, analyzing and comparing sentences of the standard answers and sentences of the student answers one by using a hybrid neural network model, and calculating the similarity of the sentences;
and S5, sequencing according to the similarity of the sentences, matching the student answers with the standard answers, analyzing the student answers according to the matching results, and giving corresponding answer guidance.
In step S1, the text data is formalized, specifically: given a Standard answer Standard, splitting the Standard answer Standard into m Standard answer sentences, wherein s1, s2, …, si, … and sm, and si represents the ith sentence in the Standard answer; the given Student answer Student splits it into { a1, a2, …, ak, …, an }, and n Student answer sentences, where ak represents the kth sentence in the Student answers.
The step S2 includes the steps of:
s201, converting words in a corpus into word vectors by using a bert model structure, enabling the words to have own meanings in space vectors, and expressing the words by using uniform dimensions;
s202, the word vectors are substituted into the sentences of the text data to obtain vector representations of the Standard answer sentences and the Student answer sentences, specifically m sentence vector representations of Standard answer Standard = { S1, S2, …, si, …, sm }, n sentence vector representations of Student answer Student = { a1, a2, …, ak, …, an }, where si represents the ith sentence in the Standard answers, and ak represents the kth sentence in the Student answers.
The step S3 includes the steps of:
s301, obtaining an individual learner 1, an individual learner 2, … and an individual learner T based on a neural network model processed by natural language;
s302, integrating the T individual learners by utilizing an integration idea to obtain a more excellent hybrid neural network analysis model.
5. The intelligent analysis answer to test questions method based on natural language processing as claimed in claim 1, wherein said step S4 comprises the steps of:
s401, carrying out permutation and combination on vector representations of the standard answer sentences and the student answer sentences to obtain sum = m × n combinations, specifically { (S1, a 1), (S1, a 2), …, (S1, an); (s 2, a 1), (s 2, a 2), …, (s 2, an); …; (sm, a 1), (sm, a 2), …, (sm, an) }; wherein sm represents the nth sentence in the standard answer, and an represents the nth sentence in the student answer;
s402, carrying out similarity analysis on the sum combination based on a mixed neural network model to obtain the matching degree of the student answers and the standard answers;
the step S5 includes the steps of:
s501, sorting the sum combinations in the step S402 according to similarity, and selecting the student answer sentences with the highest matching degree with each standard answer sentence, wherein the specific criteria are as follows:
s5011, n sentences { a1, a2, …, ak, …, an } of student answers are not reusable, where ak represents the kth sentence in the student answers;
s5012, if one student answer sentence is matched with a plurality of standard answer sentences, preferentially selecting a combination with high matching degree, and if the matching degrees are consistent, selecting according to the appearance sequence of the standard answer sentences;
s5013, if n sentences of the student answers are not enough matched with m standard answer sentences, a null matching phenomenon occurs, the standard answer sentences which do not have corresponding matching are marked as non-matching, and answers are required to be supplemented;
s5014, if n sentences of the student answers are more than m standard answer sentences, redundant matching occurs, redundant student answer sentences are marked as redundant parts, and deletion is needed;
s502, analyzing the student answers according to the matching results, giving corresponding answer guidance, specifically, judging whether the standard answer sentences are contained in the student answer sentences, and giving different conclusions according to the matching degree of each standard answer sentence and the student answer sentence, wherein the specific criteria are as follows:
s5021, if the matching degree of (si, ak) is (0.8,1.0), the student answers completely comprise standard answers, and the mark ak does not need to be improved;
s5022, if the matching degree of (si, ak) is (0.4,0.8), the student answers contain most of standard answers, and the mark ak needs to be slightly changed;
s5023, if the matching degree of the (si, ak) is (0.2,0.4 ], the student answers contain a small part of standard answers, and the mark ak needs to be changed greatly;
s5024, if the matching degree of the (si, ak) is in the range of [0.0,0.2], the student answers do not comprise standard answers, the sign si is no matching, and supplementary answers are needed; the mark ak is an excess part and needs to be deleted.
The second purpose of the invention is realized by the following technical scheme: an intelligent analysis test question answer system based on natural language processing, comprising:
the application interaction unit is an interaction unit of the server and the client; on one hand, the intelligent analysis test question answer providing system is oriented to the client and provides the function of intelligently analyzing test question answers for the client; on the other hand, the server is connected, the functions of the data preprocessing unit and the hybrid neural network unit deployed on the server are used, and the result is fed back to the user;
the data preprocessing unit is a first step for processing the input test question text data, the data processed by the data preprocessing unit enables words to have a uniform dimensional representation, and vector representation of a standard answer sentence and a student answer sentence is obtained according to the uniform dimensional representation;
the hybrid neural network unit adopts an integrated idea, combines a plurality of neural network models to form a large-scale hybrid neural network model, and is used for overcoming the defects of a single network model;
and the software and hardware resource unit is positioned at the bottom layer of the intelligent analysis test question answer system and is used for providing hardware and software support and help for the units, specifically a stored data preprocessing model and a hybrid neural network model, and a TCP/IP protocol is used for data transmission in the network transmission process.
Compared with the prior art, the invention has the following advantages and beneficial effects:
1. the invention uses the dynamic representation method of the word vector, but not the fixed representation of the word vector, solves the problem of word-polysemy, and leads the expression of the sentence vector to obtain better effect.
2. The invention uses the integrated thought and the mixed neural network model to carry out similarity analysis on the standard answer sentence and the student answer sentence, and fully utilizes the advantages and the advantages of each network model.
3. The invention gives suggestions pertinently according to different matching results of the student answer sentences and the standard answer sentences, analyzes the answers of the test questions more finely, and improves the skill of solving the questions of the students.
Drawings
FIG. 1 is a schematic flow chart of the method of the present invention
Fig. 2 is a schematic diagram of a bert structure of a dynamic characterization method of a word vector.
FIG. 3 is an integrated schematic diagram of a hybrid neural network model.
FIG. 4 is a block diagram of the system of the present invention.
Detailed Description
The present invention will be described in further detail with reference to examples and drawings, but the present invention is not limited thereto.
In this embodiment, the standard answer is the answer to the 40 th political question (1) of the national college entrance examination in 2020, which is specifically:
practice is the basis of understanding, and the correct understanding of complex things can be completed through a plurality of iterations from practice to understanding and then from the knowledge of practice; the truth is concrete, historical and an infinite process which is continuously developed. The change of the diagnosis and treatment scheme reflects that the knowledge of the new coronary pneumonia is based on the diagnosis and treatment practice, is a continuous and complete process from no deepening to deeper and incomplete to more comprehensive, and is a process for guiding the diagnosis and treatment practice and continuously receiving the examination of the diagnosis and treatment practice.
The answers of the students adopted in this embodiment are specifically:
the practice is the basis of the knowledge, the knowledge has adverse effect on the practice, the knowledge has repeatability, limitless and upgradability, and the pursuit of the truth is a process; the truth is objective, concrete and conditional, requires correct treatment for errors, and is continuously enriched, developed and perfected in practice. The change of the diagnosis and treatment scheme from the first edition to the seventh edition continuously explores a relatively complete diagnosis and treatment system on the basis of diagnosis and treatment practice, and embodies the repeatability, the limitless property and the upgradability of the cognition.
Referring to fig. 1, the intelligent answer analysis method for test questions based on natural language processing provided in this embodiment includes the following steps:
s1, acquiring text data of test questions, including standard answers and student answers, and performing formalized processing on the text data, wherein the specific process is as follows:
s101, splitting the obtained Standard answers according to semicolons and periods to obtain independent sentences, for example, given Standard answer Standard, splitting the Standard answers into m Standard answer sentences of { S1, S2, …, si, …, sm } according to semicolons and periods, wherein si represents the ith sentence of the Standard answers. In connection with the background example, three standard answer sentences are obtained as follows:
{ s1: practice is the basis of understanding, and the correct understanding of complex things can be completed through a plurality of iterations from practice to understanding and then from the knowledge of practice;
s2: the truth is concrete, historical and an infinite process which is continuously developed.
s3: the change of the diagnosis and treatment scheme reflects that the knowledge of the new coronary pneumonia is based on the diagnosis and treatment practice, is a continuous and complete process from no deepening to deeper and incomplete to more comprehensive, and is a process for guiding the diagnosis and treatment practice and continuously receiving the examination of the diagnosis and treatment practice. }
S102, splitting the obtained Student answers according to semicolons and periods to obtain independent sentences, for example, given Student answer Student can be split into n Student answer sentences of { a1, a2, …, ak, …, an } according to semicolons and periods, wherein ak represents the kth sentence of the Student answers. In connection with the background example, three student answer sentences are obtained as follows:
{ a1: the practice is the basis of the knowledge, the knowledge has adverse effect on the practice, the knowledge has repeatability, limitless and upgradability, and the pursuit of the truth is a process;
a2: the truth is objective, concrete and conditional, requires correct treatment for errors, and is continuously enriched, developed and perfected in practice.
a3: the change of the diagnosis and treatment scheme from the first edition to the seventh edition continuously explores a relatively complete diagnosis and treatment system on the basis of diagnosis and treatment practice, and embodies the repeatability, the limitless property and the upgradability of the cognition. }
S2, mapping the words into vectors with unified dimensionality for representation based on a dynamic representation method of word vectors, and substituting the vectors into sentences to obtain vector representation of each sentence, wherein the specific process is as follows:
s201, referring to the graph shown in FIG. 2, words in a corpus are converted into word vectors by using a bert model structure, so that the words have own meanings in space vectors and are represented by using uniform dimensions;
s2011, in FIG. 2, E 1 ,E 2 ,...,E N A vector representation representing a word; trm is called Transformer, is the feature extractor with the best effect at present, is proposed by Google in 2017, and is essentially self attention (self attention) overlay structure; t is 1 ,T 2 ,...,T N Representing a downstream Task (Task), and training the vector representation of the words according to the downstream Task to further obtain the vector representation of the words, namely word vectors;
s2012, in the figure 2, the bert model structure uses the bidirectional language model to pre-train the language model, so that a better effect is obtained.
S202, the word vectors are brought into the sentences of the text data to obtain vector representation of the standard answer sentences and the student answer sentences. In combination with the background material, a standard answer sentence vector { s1, s2, s3} and a student answer sentence vector { a1, a2, a3} can be obtained.
S3, establishing a hybrid neural network model by utilizing an integrated idea based on the existing neural network model processed by natural language, and referring to fig. 3, the specific process is as follows:
s301, selecting a model with excellent performance and suitability based on the existing neural network model for natural language processing, wherein in the embodiment, the individual learner comprises three neural network models which are suitable for natural language processing, namely CNN, RNN and GNN;
s302, integrating the three individual learners by using an integration idea to obtain a hybrid neural network analysis model with more excellent performance;
s4, based on the obtained vector representation of the text data, analyzing and comparing sentences of the standard answers and the students' answers one by using a mixed neural network model, calculating the similarity of the sentences, and combining background materials, wherein the specific process is as follows:
s401, performing permutation and combination on vector representations of the standard answer sentences and the student answer sentences, and combining with the background example, to obtain sum =3 × 3=9 combinations, specifically { (S1, a 1), (S1, a 2), (S1, a 3), (S2, a 1), (S2, a 2), (S2, a 3), (S3, a 1), (S3, a 2), (S3, a 3) };
s402, carrying out similarity analysis on the sum =9 combination based on a mixed neural network model to obtain the matching degree between the student answers and the standard answers, and writing the obtained result into a matrix form by assuming that the value of complete matching is 1, as follows:
s5, sequencing according to the similarity of the sentences, matching the student answers with the standard answers, analyzing the student answers according to the matching results, giving corresponding answer guidance, and combining with background materials, wherein the specific process is as follows:
s501, sorting the 9 combinations in the step S402 according to the similarity, and obtaining the following results:
selecting the student answer sentence ak with the highest matching degree with each standard answer sentence si to obtain { (s 2, a 2) =0.854 (s 1, a 1) =0.821 (s 3, a 3) =0.689};
s502, analyzing the answers of the students according to the matching results, giving corresponding answer guidance, and combining background materials according to a judgment criterion, wherein the specific process is as follows:
(s 2, a 2) =0.854, the degree of match is at (0.8,1.0 ], mark a2 need not be improved;
(s 1, a 1) =0.821, the degree of matching is at (0.8,1.0 ], no improvement is required for the mark a 1;
(s 3, a 3) =0.689, the matching degree is at (0.4,0.8 ], mark a3 needs minor change;
no matching of standard answer sentences occurs, and the student answer sentences are redundant.
Referring to fig. 4, the present embodiment also provides an intelligent analysis answer system for test questions based on natural language processing, including:
the application interaction unit is an interaction unit of the server and the client; on one hand, the intelligent analysis test question answer providing system is oriented to the client and provides the function of intelligently analyzing test question answers for the client; on the other hand, the server is connected, the functions of the data preprocessing unit and the hybrid neural network unit deployed on the server are used, and the result is fed back to the user;
the data preprocessing unit is a first step for processing the input test question text data, the data processed by the data preprocessing unit enables words to have a uniform dimensional representation, and vector representation of a standard answer sentence and a student answer sentence is obtained according to the uniform dimensional representation;
the hybrid neural network unit adopts an integrated idea, combines a plurality of neural network models to form a large-scale hybrid neural network model, and is used for overcoming the defects of a single network model;
and the software and hardware resource unit is positioned at the bottom layer of the intelligent analysis test question answer system and is used for providing hardware and software support and help for the units, specifically a stored data preprocessing model and a hybrid neural network model, and a TCP/IP protocol is used for data transmission in the network transmission process.
The above embodiments are preferred embodiments of the present invention, but the present invention is not limited to the above embodiments, and any other changes, modifications, substitutions, combinations, and simplifications which do not depart from the spirit and principle of the present invention should be construed as equivalents thereof, and all such changes, modifications, substitutions, combinations, and simplifications are intended to be included in the scope of the present invention.
Claims (1)
1. An intelligent analysis test question answer method based on natural language processing is characterized by comprising the following steps:
s1, acquiring text data of test questions, including standard answers and student answers, and performing formalized processing on the text data;
s2, mapping words of text data into vectors with unified dimensionality for representation based on a dynamic representation method of word vectors, and substituting the vectors into sentences to obtain vector representation of each sentence, wherein the method comprises the following steps:
s201, converting words in a corpus into word vectors by using a bert model structure, enabling the words to have own meanings in space vectors, and expressing the words by using uniform dimensions;
s202, substituting word vectors into sentences of text data to obtain vector representations of Standard answer sentences and Student answer sentences, specifically to obtain m sentence vector representations of Standard answer Standard = { S1, S2, …, si, …, sm }, n sentence vector representations of Student answer Student = { a1, a2, …, ak, …, an }, where si represents the ith sentence in the Standard answers, and ak represents the kth sentence in the Student answers;
s3, establishing a hybrid neural network model by utilizing an integrated thought based on the neural network model processed by the natural language, and comprising the following steps of:
s301, obtaining an individual learner 1, an individual learner 2, … and an individual learner T based on a neural network model processed by natural language;
s302, integrating the T individual learners by utilizing an integration idea to obtain a more excellent hybrid neural network analysis model;
s4, based on the obtained vector representation of the text data, analyzing and comparing sentences of the standard answers and the students' answers one by using a mixed neural network model, and calculating the similarity of the sentences, wherein the method comprises the following steps:
s401, carrying out permutation and combination on vector representations of the standard answer sentences and the student answer sentences to obtain sum = m × n combinations, specifically { (S1, a 1), (S1, a 2), …, (S1, an); (s 2, a 1), (s 2, a 2), …, (s 2, an); …; (sm, a 1), (sm, a 2), …, (sm, an) }; wherein sm represents the nth sentence in the standard answer, and an represents the nth sentence in the student answer;
s402, carrying out similarity analysis on the sum combination based on a mixed neural network model to obtain the matching degree of the student answers and the standard answers;
s5, sequencing according to the similarity of sentences, matching the student answers with the standard answers, analyzing the student answers according to the matching results, and giving corresponding answer guidance, wherein the method comprises the following steps:
s501, sorting the sum combinations in the step S402 according to similarity, and selecting the student answer sentences with the highest matching degree with each standard answer sentence, wherein the specific criteria are as follows:
s5011, n sentences { a1, a2, …, ak, …, an } of student answers are not reusable, where ak represents the kth sentence in the student answers;
s5012, if one student answer sentence is matched with a plurality of standard answer sentences, preferentially selecting a combination with high matching degree, and if the matching degrees are consistent, selecting according to the appearance sequence of the standard answer sentences;
s5013, if n sentences of the student answers are not enough matched with m standard answer sentences, a null matching phenomenon occurs, the standard answer sentences which do not have corresponding matching are marked as non-matching, and answers are required to be supplemented;
s5014, if n sentences of the student answers are more than m standard answer sentences, redundant matching occurs, redundant student answer sentences are marked as redundant parts, and deletion is needed;
s502, analyzing the student answers according to the matching results, giving corresponding answer guidance, specifically, judging whether the standard answer sentences are contained in the student answer sentences, and giving different conclusions according to the matching degree of each standard answer sentence and the student answer sentence, wherein the specific criteria are as follows:
s5021, if the matching degree of (si, ak) is (0.8, 1.0], the student answers completely comprise standard answers, and the mark ak is no need to be improved, wherein si represents the ith sentence in the standard answers;
s5022, if the matching degree of (si, ak) is (0.4, 0.8), the student answers contain most standard answers, and the mark ak needs to be slightly changed;
s5023, if the matching degree of (si, ak) is (0.2, 0.4), the student answers contain a small part of standard answers, and the mark ak needs to be changed greatly;
s5024, if the matching degree of (si, ak) is [0.0,0.2], the student answers do not comprise standard answers, si is marked as no matching, and supplementary answers are needed; the mark ak is an excess part and needs to be deleted.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110545942.8A CN113312909B (en) | 2021-05-19 | 2021-05-19 | Intelligent analysis test question answer method and system based on natural language processing |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110545942.8A CN113312909B (en) | 2021-05-19 | 2021-05-19 | Intelligent analysis test question answer method and system based on natural language processing |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113312909A CN113312909A (en) | 2021-08-27 |
| CN113312909B true CN113312909B (en) | 2023-01-06 |
Family
ID=77373907
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110545942.8A Active CN113312909B (en) | 2021-05-19 | 2021-05-19 | Intelligent analysis test question answer method and system based on natural language processing |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113312909B (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109213999A (en) * | 2018-08-20 | 2019-01-15 | 成都佳发安泰教育科技股份有限公司 | A kind of subjective item methods of marking |
| CN110134925A (en) * | 2019-05-15 | 2019-08-16 | 北京信息科技大学 | A kind of Chinese patent text similarity calculating method |
| CN110633359A (en) * | 2019-09-04 | 2019-12-31 | 北京百分点信息科技有限公司 | Sentence equivalence judgment method and device |
| CN111325029A (en) * | 2020-02-21 | 2020-06-23 | 河海大学 | Text similarity calculation method based on deep learning integration model |
| CN112308370A (en) * | 2020-09-16 | 2021-02-02 | 湘潭大学 | Transform-based subjective question automatic scoring technology for thinking administration courses |
Family Cites Families (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10489712B2 (en) * | 2016-02-26 | 2019-11-26 | Oath Inc. | Quality-based scoring and inhibiting of user-generated content |
| US10255273B2 (en) * | 2017-06-15 | 2019-04-09 | Microsoft Technology Licensing, Llc | Method and system for ranking and summarizing natural language passages |
| CN107633721A (en) * | 2017-08-25 | 2018-01-26 | 刘宸瑜 | A kind of error analysis method, apparatus and system for language learning |
| CN108764074B (en) * | 2018-05-14 | 2019-03-19 | 山东师范大学 | Subjective item intelligently reading method, system and storage medium based on deep learning |
| CN109947912B (en) * | 2019-01-25 | 2020-06-23 | 四川大学 | Model method based on intra-paragraph reasoning and joint question answer matching |
| CN110046244B (en) * | 2019-04-24 | 2021-06-08 | 中国人民解放军国防科技大学 | Answer selection method for question-answering system |
-
2021
- 2021-05-19 CN CN202110545942.8A patent/CN113312909B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109213999A (en) * | 2018-08-20 | 2019-01-15 | 成都佳发安泰教育科技股份有限公司 | A kind of subjective item methods of marking |
| CN110134925A (en) * | 2019-05-15 | 2019-08-16 | 北京信息科技大学 | A kind of Chinese patent text similarity calculating method |
| CN110633359A (en) * | 2019-09-04 | 2019-12-31 | 北京百分点信息科技有限公司 | Sentence equivalence judgment method and device |
| CN111325029A (en) * | 2020-02-21 | 2020-06-23 | 河海大学 | Text similarity calculation method based on deep learning integration model |
| CN112308370A (en) * | 2020-09-16 | 2021-02-02 | 湘潭大学 | Transform-based subjective question automatic scoring technology for thinking administration courses |
Non-Patent Citations (2)
| Title |
|---|
| 何超.主观题自动评分模型研究与验证.《中国优秀硕士学位论文全文数据库信息科技辑》.2017,I138-1211. * |
| 王冲.短答案自动评分方法研究.《中国优秀硕士学位论文全文数据库信息科技辑》.2021,(第1期),I138-2626. * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113312909A (en) | 2021-08-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Mathew et al. | NLP-based personal learning assistant for school education | |
| Jiang | Application of artificial intelligence technology in music education supported by wireless network | |
| Zhu et al. | Construction and analysis of intelligent english teaching model assisted by personalized virtual corpus by big data analysis | |
| Ahmed et al. | Enhancing The Teaching And Learning Of English For Specific Purposes (Esp) With Chatgpt | |
| CN109119064A (en) | A kind of implementation method suitable for overturning the Oral English Teaching system in classroom | |
| CN113312909B (en) | Intelligent analysis test question answer method and system based on natural language processing | |
| Kazachkova et al. | The Use of Digital Resources as a Means of Forming a Foreign Language Communicative Competence of University Students (on the Example of Teaching Special Vocabulary) | |
| Jing | A teaching system of English online course based on artificial intelligence | |
| Andrejević et al. | E-LEARNING WEB APPLICATION FOR TEACHING ENGLISH FOR SPECIFIC PURPOSES-“ENGLISH AT A CLICK” | |
| CN112015920A (en) | Intelligent auxiliary learning system based on knowledge graph and edge calculation | |
| Han et al. | A study on exploring the path of psychology and civics teaching reform in universities based on artificial intelligence | |
| Zhai | Research on Teaching Mode of College English Translating Course from the Perspective of Eco-translatology | |
| CN114997150B (en) | Method and device for researching and establishing grammar knowledge resources based on international Chinese education and teaching | |
| Yuan | Research on college english teaching mode based on artificial intelligence | |
| Liu | Research on the model of college English network learning platform based on education big data | |
| Lulu et al. | Integrated forestry English education network platform based on artificial intelligence | |
| Wang | Optimization and Evaluation of Oral English CAF Based on Artificial Intelligence and Corpus | |
| Ding | Development of computer-aided English listening system based on BS architecture | |
| Aleedy et al. | Towards deep learning-powered chatbot for translation learning | |
| Zhang et al. | Teacher Being The Leading Role: Introducing Comparative Literature Mechanism In Literature Translation Course | |
| Zhao et al. | Design and Implementation of a Teaching Verbal Behavior Analysis Aid in Instructional Videos | |
| Dong et al. | The Application of Artificial Intelligence in Computer-Assisted College English Teaching Tasks | |
| Liu | Informatization of Constructive English Learning Platform Based on Improved GLR Algorithm | |
| Yang et al. | In-Depth Application of Artificial Intelligence-Generated Content AIGC Large Model in Higher Education | |
| Lou | Study on vocabulary learning strategies for Chinese English-Majors |
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 |