CN104572617A - Oral test answer deviation detection method and device - Google Patents
Oral test answer deviation detection method and device Download PDFInfo
- Publication number
- CN104572617A CN104572617A CN201410841199.0A CN201410841199A CN104572617A CN 104572617 A CN104572617 A CN 104572617A CN 201410841199 A CN201410841199 A CN 201410841199A CN 104572617 A CN104572617 A CN 104572617A
- Authority
- CN
- China
- Prior art keywords
- examinee
- answer
- feature
- semantic
- semantic feature
- 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
Landscapes
- Machine Translation (AREA)
Abstract
The invention discloses an oral test answer deviation detection method and device and belongs to the technical field of speech data processing. The method includes: recognizing an answer text sequence of an examinee according to an answer audio file and corresponding examination question information of the examinee; performing semantic analysis through the answer text sequence of the examinee, and extracting semantic features of the examinee; performing part-of-speech tagging on the answer text sequence of the examinee, generating a syntax tree of answer test, and extracting grammatical features of the examinee according to syntax tree features; detecting whether answers of the examinee deviate from the examination question range or not according to the semantic features and the grammatical features of the examinee and examination question information. The semantic features and the grammatical features of the answer audio file of the examinee are extracted and taken as bases for objectively detecting whether the answers of the examinee deviate from the examination question range or not, powerful help is provided for scoring of oral test, and fairness and accuracy of the oral test are improved.
Description
Technical field
The present invention relates to language data process technical field, particularly a kind of SET tricky question detection method and device.
Background technology
In recent years along with the aggravation of socioeconomic development and the trend of globalization, the enthusiasm of people to language learning reaches unprecedented height.Corresponding as detecting quality of instruction, checking the evaluation and test of the language of results of learning to it is also proposed more and more higher requirement to the assess effectiveness that assessment objectivity, fairness and scale are tested.
Existing computing machine automatic scoring technology more attention is the correlated characteristic of the phonetics aspect of examination, and so, examinee only needs pronunciation fluent and clear in answer to a question, and points-scoring system all can provide certain mark., have the poor examinee of some abilities and skillfully recite some conventional texts in advance, also can obtain a mark from system, the fairness causing SET to be marked is had a strong impact on.
Summary of the invention
In order to solve the problem, embodiments provide a kind of SET tricky question detection method and device.Described technical scheme is as follows:
On the one hand, provide a kind of SET tricky question detection method, described method comprises:
According to answer audio file and the corresponding examination question information of examinee, identify the answer text sequence of examinee;
By carrying out semantic analysis to described examinee's answer text sequence, extract the semantic feature of examinee;
Part-of-speech tagging is carried out to described examinee's answer text sequence, generates the syntax tree of answer text, according to described syntax tree feature, extract the grammar property of examinee;
According to the semantic feature of described examinee, the grammar property of described examinee and examination question information, whether the answer detecting described examinee departs from examination paper scope.
Alternatively, according to answer audio file and the corresponding examination question information of examinee, the answer text sequence identifying examinee comprises:
According to answer audio file and the corresponding examination question information of examinee, obtain the language feature of examinee, and generate keywords database and thesaurus by language parse;
According to the language feature of described examinee, utilize preset language model and acoustic model, identify the answer text sequence of examinee.
Alternatively, by carrying out semantic analysis to described examinee's answer text sequence, the semantic feature extracting examinee comprises:
Semantic analysis is carried out to the answer text sequence of described examinee, extract the semantic feature of examinee, the semantic feature of described examinee at least comprises: the keyword distribution characteristics of the answer text of the examinee calculated based on described keywords database and described thesaurus, and the text similarity feature utilizing term vector method to calculate.
Alternatively, the keyword distribution characteristics calculating the answer text of examinee based on described keywords database and described thesaurus comprises:
Based on described keywords database and described thesaurus, according to vector space model, calculate keyword vector;
Based on described keywords database and described thesaurus, in conjunction with the answer text series of described examinee, by both quantitative proportions as keyword distribution characteristics.
Alternatively, utilize term vector method to calculate text similarity feature to comprise:
Based on described keyword and nearly justice, calculate the TF-IDF vector characteristic of examinee's answer, recycling Shallow Semantic Parsing model extraction semantic feature vector, uses described semantic feature to calculate the answer text similarity eigenvector of examinee;
Using examination question information and examinee's answer audio file, set up word2vec model, obtain the term vector of examination paper and examinee's answer, by calculating the distance between term vector, obtaining the text similarity feature of examinee's answer.
Alternatively, according to the semantic feature of described examinee, the grammar property of described examinee and examination question information, whether the answer detecting described examinee departs from examination paper scope comprises:
Utilize multiple linear regression analysis method, according to the semantic feature of described examinee, the grammar property of described examinee and examination question information, calculate the answer scope of examinee; When the examinee's answer scope calculated is greater than preset value, think that the answer of examinee departs from examination paper scope; When the examinee's answer scope calculated is less than preset value, think that the answer of examinee does not depart from examination paper scope.
On the other hand, provide a kind of SET tricky question pick-up unit, described device comprises:
Text sequence identification module, for according to the answer audio file of examinee and corresponding examination question information, identifies the answer text sequence of examinee;
Semantic feature extraction module, for by carrying out semantic analysis to described examinee's answer text sequence, extracts the semantic feature of examinee;
Grammar property extraction module, carries out part-of-speech tagging for described examinee's answer text sequence, generates the syntax tree of answer text, according to described syntax tree feature, extracts the grammar property of examinee;
Tricky question judge module, for grammar property and the examination question information of the semantic feature according to described examinee, described examinee, whether the answer detecting described examinee departs from examination paper scope.
Alternatively, described text sequence identification module is used for, according to the answer audio file of examinee and corresponding examination question information, obtaining the language feature of examinee, and generates keywords database and thesaurus by language parse; According to the language feature of described examinee, utilize preset language model and acoustic model, identify the answer text sequence of examinee.
Alternatively, described semantic feature extraction module is used for carrying out semantic analysis to the answer text sequence of described examinee, extract the semantic feature of examinee, the semantic feature of described examinee at least comprises: the keyword distribution characteristics of the answer text of the examinee calculated based on described keywords database and described thesaurus, and the text similarity feature utilizing term vector device to calculate; The keyword distribution characteristics calculating the answer text of examinee based on described keywords database and described thesaurus comprises: based on described keywords database and described thesaurus, according to vector space model, calculates keyword vector; Based on described keywords database and described thesaurus, in conjunction with the answer text series of described examinee, by both quantitative proportions as keyword distribution characteristics; Utilize term vector device to calculate text similarity feature to comprise: based on described keyword and nearly justice, calculate the TF-IDF vector characteristic of examinee's answer, recycling Shallow Semantic Parsing model extraction semantic feature vector, uses described semantic feature to calculate the answer text similarity eigenvector of examinee; Using examination question information and examinee's answer audio file, set up word2vec model, obtain the term vector of examination paper and examinee's answer, by calculating the distance between term vector, obtaining the text similarity feature of examinee's answer.
Alternatively, described tricky question judge module is used for utilizing multiple linear regression device, according to the semantic feature of described examinee, the grammar property of described examinee and examination question information, calculates the answer scope of examinee; When the examinee's answer scope calculated is greater than preset value, think that the answer of examinee departs from examination paper scope; When the examinee's answer scope calculated is less than preset value, think that the answer of examinee does not depart from examination paper scope.
The beneficial effect that the technical scheme that the embodiment of the present invention provides is brought is:
By extracting semantic feature and the grammar property of examinee's answer audio file, and be characterized as basis with these and detect examinee's answer objectively and whether depart from examination paper scope, scoring for SET provides strong help, improves fairness and the accuracy of SET.
Accompanying drawing explanation
In order to be illustrated more clearly in the technical scheme in the embodiment of the present invention, below the accompanying drawing used required in describing embodiment is briefly described, apparently, accompanying drawing in the following describes is only some embodiments of the present invention, for those of ordinary skill in the art, under the prerequisite not paying creative work, other accompanying drawing can also be obtained according to these accompanying drawings.
Fig. 1 is the SET tricky question detection method process flow diagram that the embodiment of the present invention provides;
Fig. 2 is the SET tricky question structure of the detecting device schematic diagram that the embodiment of the present invention provides.
Embodiment
For making the object, technical solutions and advantages of the present invention clearly, below in conjunction with accompanying drawing, embodiment of the present invention is described further in detail.
Fig. 1 is the process flow diagram of the SET tricky question detection method that the embodiment of the present invention provides.See Fig. 1, the method comprises:
101, according to answer audio file and the corresponding examination question information of examinee, the answer text sequence of examinee is identified;
Wherein, the answer text sequence of this examinee can comprise word level sequence and phoneme level sequence etc.
In embodiments of the present invention, according to answer audio file and the corresponding examination question information of examinee, the answer text sequence identifying examinee comprises: according to answer audio file and the corresponding examination question information of examinee, obtains the language feature of examinee, and generates keywords database and thesaurus by language parse; According to the language feature of this examinee, utilize preset language model and acoustic model, identify the answer text sequence of examinee.
102, by carrying out semantic analysis to this examinee's answer text sequence, the semantic feature of examinee is extracted;
In embodiments of the present invention, by carrying out semantic analysis to this examinee's answer text sequence, the semantic feature extracting examinee comprises: carry out semantic analysis to the answer text sequence of this examinee, extract the semantic feature of examinee, the semantic feature of this examinee at least comprises: based on the keyword distribution characteristics of this keywords database with the answer text of the examinee of this thesaurus calculating, and the text similarity feature utilizing term vector method to calculate.
Wherein, in embodiments of the present invention, the keyword distribution characteristics calculating the answer text of examinee based on this keywords database and this thesaurus comprises: based on this keywords database and this thesaurus, according to vector space model, calculates keyword vector; Based on this keywords database and this thesaurus, in conjunction with the answer text series of this examinee, by both quantitative proportions as keyword distribution characteristics.
Wherein, in embodiments of the present invention, utilize term vector method to calculate text similarity feature to comprise: based on this keyword and nearly justice, calculate the TF-IDF vector characteristic of examinee's answer, recycling Shallow Semantic Parsing model extraction semantic feature vector, uses this semantic feature to calculate the answer text similarity eigenvector of examinee; Using examination question information and examinee's answer audio file, set up word2vec model, obtain the term vector of examination paper and examinee's answer, by calculating the distance between term vector, obtaining the text similarity feature of examinee's answer.
103, part-of-speech tagging is carried out to this examinee's answer text sequence, generate the syntax tree of answer text, according to this syntax tree feature, extract the grammar property of examinee;
In embodiments of the present invention, part-of-speech tagging method can being utilized, use the context free grammar method based on probability, generate the syntax tree of answer text, according to features such as the shape sizes of syntax tree, extracting the grammar property for detecting answer.
104, according to the semantic feature of this examinee, the grammar property of this examinee and examination question information, whether the answer detecting this examinee departs from examination paper scope.
In embodiments of the present invention, utilize multiple linear regression analysis method, according to the semantic feature of this examinee, the grammar property of this examinee and examination question information, calculate the answer scope of examinee; When the examinee's answer scope calculated is greater than preset value, think that the answer of examinee departs from examination paper scope; When the examinee's answer scope calculated is less than preset value, think that the answer of examinee does not depart from examination paper scope.
The method that the embodiment of the present invention provides, by the semantic feature by extracting examinee's answer audio file and grammar property, and be characterized as basis with these and detect examinee's answer objectively and whether depart from examination paper scope, scoring for SET provides strong help, improves fairness and the accuracy of SET.
Fig. 2 is the structural representation of the SET tricky question pick-up unit that the embodiment of the present invention provides.See Fig. 2, this device comprises text sequence identification module 21, semantic feature extraction module 22, grammar property extraction module 23 and tricky question judge module 24.Wherein,
Text sequence identification module 21, for according to the answer audio file of examinee and corresponding examination question information, identifies the answer text sequence of examinee; Text sequence identification module 21 is connected with semantic feature extraction module 22, and semantic feature extraction module 22, for by carrying out semantic analysis to this examinee's answer text sequence, extracts the semantic feature of examinee; Text sequence identification module 21 is connected with grammar property extraction module 23, and grammar property extraction module 23 carries out part-of-speech tagging for this examinee's answer text sequence, generates the syntax tree of answer text, according to this syntax tree feature, extracts the grammar property of examinee; Semantic feature extraction module 22 is connected with tricky question judge module 24 with grammar property extraction module 23, tricky question judge module 24 is for according to the semantic feature of this examinee, the grammar property of this examinee and examination question information, and whether the answer detecting this examinee departs from examination paper scope.
Alternatively, text recognition sequence module 21, for according to the answer audio file of examinee and corresponding examination question information, obtains the language feature of examinee, and generates keywords database and thesaurus by language parse; According to the language feature of this examinee, utilize preset language model and acoustic model, identify the answer text sequence of examinee.
Alternatively, this semantic feature extraction module 22 is for carrying out semantic analysis to the answer text sequence of this examinee, extract the semantic feature of examinee, the semantic feature of this examinee at least comprises: based on the keyword distribution characteristics of this keywords database with the answer text of the examinee of this thesaurus calculating, and the text similarity feature utilizing term vector device to calculate; The keyword distribution characteristics calculating the answer text of examinee based on this keywords database and this thesaurus comprises: based on this keywords database and this thesaurus, according to vector space model, calculates keyword vector; Based on this keywords database and this thesaurus, in conjunction with the answer text series of this examinee, by both quantitative proportions as keyword distribution characteristics; Utilize term vector device to calculate text similarity feature to comprise: based on this keyword and nearly justice, calculate the TF-IDF vector characteristic of examinee's answer, recycling Shallow Semantic Parsing model extraction semantic feature vector, uses this semantic feature to calculate the answer text similarity eigenvector of examinee;
Using examination question information and examinee's answer audio file, set up word2vec model, obtain the term vector of examination paper and examinee's answer, by calculating the distance between term vector, obtaining the text similarity feature of examinee's answer.
Alternatively, this tricky question judge module 24, for utilizing multiple linear regression device, according to the semantic feature of this examinee, the grammar property of this examinee and examination question information, calculates the answer scope of examinee; When the examinee's answer scope calculated is greater than preset value, think that the answer of examinee departs from examination paper scope; When the examinee's answer scope calculated is less than preset value, think that the answer of examinee does not depart from examination paper scope.
The device that the embodiment of the present invention provides, by extracting semantic feature and the grammar property of examinee's answer audio file, and be characterized as basis with these and detect examinee's answer objectively and whether depart from examination paper scope, scoring for SET provides strong help, improves fairness and the accuracy of SET.
One of ordinary skill in the art will appreciate that all or part of step realizing above-described embodiment can have been come by hardware, the hardware that also can carry out instruction relevant by program completes, this program can be stored in a kind of computer-readable recording medium, the above-mentioned storage medium mentioned can be ROM (read-only memory), disk or CD etc.
The foregoing is only preferred embodiment of the present invention, not in order to limit the present invention, within the spirit and principles in the present invention all, any amendment done, equivalent replacement, improvement etc., all should be included within protection scope of the present invention.
Claims (10)
1. a SET tricky question detection method, is characterized in that, described method comprises:
According to answer audio file and the corresponding examination question information of examinee, identify the answer text sequence of examinee;
By carrying out semantic analysis to described examinee's answer text sequence, extract the semantic feature of examinee;
Part-of-speech tagging is carried out to described examinee's answer text sequence, generates the syntax tree of answer text, according to described syntax tree feature, extract the grammar property of examinee;
According to the semantic feature of described examinee, the grammar property of described examinee and examination question information, whether the answer detecting described examinee departs from examination paper scope.
2. method according to claim 1, is characterized in that, according to answer audio file and the corresponding examination question information of examinee, the answer text sequence identifying examinee comprises:
According to answer audio file and the corresponding examination question information of examinee, obtain the language feature of examinee, and generate keywords database and thesaurus by language parse;
According to the language feature of described examinee, utilize preset language model and acoustic model, identify the answer text sequence of examinee.
3. method according to claim 1, is characterized in that, by carrying out semantic analysis to described examinee's answer text sequence, the semantic feature extracting examinee comprises:
Semantic analysis is carried out to the answer text sequence of described examinee, extract the semantic feature of examinee, the semantic feature of described examinee at least comprises: the keyword distribution characteristics of the answer text of the examinee calculated based on described keywords database and described thesaurus, and the text similarity feature utilizing term vector method to calculate.
4. method according to claim 3, is characterized in that, the keyword distribution characteristics calculating the answer text of examinee based on described keywords database and described thesaurus comprises:
Based on described keywords database and described thesaurus, according to vector space model, calculate keyword vector;
Based on described keywords database and described thesaurus, in conjunction with the answer text series of described examinee, by both quantitative proportions as keyword distribution characteristics.
5. method according to claim 3, is characterized in that, utilizes term vector method to calculate text similarity feature and comprises:
Based on described keyword and nearly justice, calculate the TF-IDF vector characteristic of examinee's answer, recycling Shallow Semantic Parsing model extraction semantic feature vector, uses described semantic feature to calculate the answer text similarity eigenvector of examinee;
Using examination question information and examinee's answer audio file, set up word2vec model, obtain the term vector of examination paper and examinee's answer, by calculating the distance between term vector, obtaining the text similarity feature of examinee's answer.
6. method according to claim 1, is characterized in that, according to the semantic feature of described examinee, the grammar property of described examinee and examination question information, whether the answer detecting described examinee departs from examination paper scope comprises:
Utilize multiple linear regression analysis method, according to the semantic feature of described examinee, the grammar property of described examinee and examination question information, calculate the answer scope of examinee; When the examinee's answer scope calculated is greater than preset value, think that the answer of examinee departs from examination paper scope; When the examinee's answer scope calculated is less than preset value, think that the answer of examinee does not depart from examination paper scope.
7. a SET tricky question pick-up unit, is characterized in that, described device comprises:
Text sequence identification module, for according to the answer audio file of examinee and corresponding examination question information, identifies the answer text sequence of examinee;
Semantic feature extraction module, for by carrying out semantic analysis to described examinee's answer text sequence, extracts the semantic feature of examinee;
Grammar property extraction module, carries out part-of-speech tagging for described examinee's answer text sequence, generates the syntax tree of answer text, according to described syntax tree feature, extracts the grammar property of examinee;
Tricky question judge module, for grammar property and the examination question information of the semantic feature according to described examinee, described examinee, whether the answer detecting described examinee departs from examination paper scope.
8. device according to claim 7, is characterized in that, described text sequence identification module is used for, according to the answer audio file of examinee and corresponding examination question information, obtaining the language feature of examinee, and generates keywords database and thesaurus by language parse; According to the language feature of described examinee, utilize preset language model and acoustic model, identify the answer text sequence of examinee.
9. device according to claim 7, it is characterized in that, described semantic feature extraction module is used for carrying out semantic analysis to the answer text sequence of described examinee, extract the semantic feature of examinee, the semantic feature of described examinee at least comprises: the keyword distribution characteristics of the answer text of the examinee calculated based on described keywords database and described thesaurus, and the text similarity feature utilizing term vector device to calculate; The keyword distribution characteristics calculating the answer text of examinee based on described keywords database and described thesaurus comprises: based on described keywords database and described thesaurus, according to vector space model, calculates keyword vector; Based on described keywords database and described thesaurus, in conjunction with the answer text series of described examinee, by both quantitative proportions as keyword distribution characteristics; Utilize term vector device to calculate text similarity feature to comprise: based on described keyword and nearly justice, calculate the TF-IDF vector characteristic of examinee's answer, recycling Shallow Semantic Parsing model extraction semantic feature vector, uses described semantic feature to calculate the answer text similarity eigenvector of examinee; Using examination question information and examinee's answer audio file, set up word2vec model, obtain the term vector of examination paper and examinee's answer, by calculating the distance between term vector, obtaining the text similarity feature of examinee's answer.
10. device according to claim 7, is characterized in that, described tricky question judge module is used for utilizing multiple linear regression device, according to the semantic feature of described examinee, the grammar property of described examinee and examination question information, calculates the answer scope of examinee; When the examinee's answer scope calculated is greater than preset value, think that the answer of examinee departs from examination paper scope; When the examinee's answer scope calculated is less than preset value, think that the answer of examinee does not depart from examination paper scope.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410841199.0A CN104572617A (en) | 2014-12-30 | 2014-12-30 | Oral test answer deviation detection method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410841199.0A CN104572617A (en) | 2014-12-30 | 2014-12-30 | Oral test answer deviation detection method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN104572617A true CN104572617A (en) | 2015-04-29 |
Family
ID=53088721
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410841199.0A Pending CN104572617A (en) | 2014-12-30 | 2014-12-30 | Oral test answer deviation detection method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104572617A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106126613A (en) * | 2016-06-22 | 2016-11-16 | 苏州大学 | One composition of digressing from the subject determines method and device |
| CN106294505A (en) * | 2015-06-10 | 2017-01-04 | 华中师范大学 | A kind of method and apparatus feeding back answer |
| CN107818795A (en) * | 2017-11-15 | 2018-03-20 | 苏州驰声信息科技有限公司 | The assessment method and device of a kind of Oral English Practice |
| CN108399061A (en) * | 2018-02-26 | 2018-08-14 | 浙江创课教育科技有限公司 | The control method and system of playout of voice |
| CN110287396A (en) * | 2019-05-07 | 2019-09-27 | 清华大学 | Text matching technique and device |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102831558A (en) * | 2012-07-20 | 2012-12-19 | 桂林电子科技大学 | System and method for automatically scoring college English compositions independent of manual pre-scoring |
| CN103761975A (en) * | 2014-01-07 | 2014-04-30 | 苏州思必驰信息科技有限公司 | Method and device for oral evaluation |
-
2014
- 2014-12-30 CN CN201410841199.0A patent/CN104572617A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102831558A (en) * | 2012-07-20 | 2012-12-19 | 桂林电子科技大学 | System and method for automatically scoring college English compositions independent of manual pre-scoring |
| CN103761975A (en) * | 2014-01-07 | 2014-04-30 | 苏州思必驰信息科技有限公司 | Method and device for oral evaluation |
Non-Patent Citations (3)
| Title |
|---|
| 宋阳 等: "一种自发性口语评测文本语义相似度评分特征提取方法", 《集成技术》 * |
| 李艳玲 等: "多特征融合的英语口语考试自动评分***的研究", 《电子与信息学报》 * |
| 汪文棣 等: "英语口语机考评分***研究与实现", 《电子器件》 * |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106294505A (en) * | 2015-06-10 | 2017-01-04 | 华中师范大学 | A kind of method and apparatus feeding back answer |
| CN106294505B (en) * | 2015-06-10 | 2020-07-07 | 华中师范大学 | Answer feedback method and device |
| CN106126613A (en) * | 2016-06-22 | 2016-11-16 | 苏州大学 | One composition of digressing from the subject determines method and device |
| CN107818795A (en) * | 2017-11-15 | 2018-03-20 | 苏州驰声信息科技有限公司 | The assessment method and device of a kind of Oral English Practice |
| CN107818795B (en) * | 2017-11-15 | 2020-11-17 | 苏州驰声信息科技有限公司 | Method and device for evaluating oral English |
| CN108399061A (en) * | 2018-02-26 | 2018-08-14 | 浙江创课教育科技有限公司 | The control method and system of playout of voice |
| CN110287396A (en) * | 2019-05-07 | 2019-09-27 | 清华大学 | Text matching technique and device |
| CN110287396B (en) * | 2019-05-07 | 2021-08-03 | 清华大学 | Text matching method and device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101727903B (en) | Pronunciation quality assessment and error detection method based on fusion of multiple characteristics and multiple systems | |
| CN101751919B (en) | Spoken Chinese stress automatic detection method | |
| US8990082B2 (en) | Non-scorable response filters for speech scoring systems | |
| US9704413B2 (en) | Non-scorable response filters for speech scoring systems | |
| US9514109B2 (en) | Computer-implemented systems and methods for scoring of spoken responses based on part of speech patterns | |
| CN102034475B (en) | Method for interactively scoring open short conversation by using computer | |
| CN104572617A (en) | Oral test answer deviation detection method and device | |
| CN107305541A (en) | Speech recognition text segmentation method and device | |
| CN105741831A (en) | Spoken language evaluation method based on grammatical analysis and spoken language evaluation system | |
| Snover et al. | A lexically-driven algorithm for disfluency detection | |
| CN103761975A (en) | Method and device for oral evaluation | |
| CN109192194A (en) | Voice data mask method, device, computer equipment and storage medium | |
| CN109686383A (en) | A kind of speech analysis method, device and storage medium | |
| US20120245942A1 (en) | Computer-Implemented Systems and Methods for Evaluating Prosodic Features of Speech | |
| CN110415725B (en) | Method and system for evaluating pronunciation quality of second language using first language data | |
| CN104464755A (en) | Voice evaluation method and device | |
| CN109166569B (en) | Detection method and device for phoneme mislabeling | |
| CN109614623A (en) | Composition processing method and system based on syntactic analysis | |
| CN109346108B (en) | Operation checking method and system | |
| CN111427996B (en) | Method and device for extracting date and time from man-machine interaction text | |
| CN113822052A (en) | Text error detection method and device, electronic equipment and storage medium | |
| CN113205729A (en) | Foreign student-oriented speech evaluation method, device and system | |
| WO2019075827A1 (en) | Voice evaluation method and device | |
| CN112116181B (en) | Classroom quality model training method, classroom quality evaluation method and classroom quality evaluation device | |
| US8666987B2 (en) | Apparatus and method for processing documents to extract expressions and descriptions |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20150429 |
|
| RJ01 | Rejection of invention patent application after publication |