CN112765351A - Deep learning-fused student text feedback fine-grained analysis device and method - Google Patents
Deep learning-fused student text feedback fine-grained analysis device and method Download PDFInfo
- Publication number
- CN112765351A CN112765351A CN202110059701.2A CN202110059701A CN112765351A CN 112765351 A CN112765351 A CN 112765351A CN 202110059701 A CN202110059701 A CN 202110059701A CN 112765351 A CN112765351 A CN 112765351A
- Authority
- CN
- China
- Prior art keywords
- text
- data
- deep learning
- word
- student
- 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
- 238000000034 method Methods 0.000 title claims abstract description 29
- 238000004458 analytical method Methods 0.000 title claims description 35
- 239000013598 vector Substances 0.000 claims abstract description 46
- 238000013135 deep learning Methods 0.000 claims abstract description 34
- 230000008451 emotion Effects 0.000 claims abstract description 19
- 238000012545 processing Methods 0.000 claims abstract description 12
- 238000013527 convolutional neural network Methods 0.000 claims abstract description 10
- 230000002457 bidirectional effect Effects 0.000 claims abstract description 8
- 230000004927 fusion Effects 0.000 claims abstract description 8
- 238000011156 evaluation Methods 0.000 claims description 20
- 238000012549 training Methods 0.000 claims description 13
- 230000007935 neutral effect Effects 0.000 claims description 6
- 238000007781 pre-processing Methods 0.000 claims description 6
- 238000012805 post-processing Methods 0.000 claims description 5
- 230000006870 function Effects 0.000 claims description 4
- 238000011478 gradient descent method Methods 0.000 claims description 4
- 238000012015 optical character recognition Methods 0.000 claims description 4
- 238000012706 support-vector machine Methods 0.000 claims description 4
- 230000001419 dependent effect Effects 0.000 claims description 3
- 230000008569 process Effects 0.000 claims description 2
- 238000010224 classification analysis Methods 0.000 abstract description 2
- 238000004891 communication Methods 0.000 description 5
- 230000000694 effects Effects 0.000 description 5
- 230000008901 benefit Effects 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000004140 cleaning Methods 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000009193 crawling Effects 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000013136 deep learning model Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 239000003814 drug Substances 0.000 description 1
- 230000002996 emotional effect Effects 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 230000002452 interceptive effect Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000033764 rhythmic process Effects 0.000 description 1
- 208000024891 symptom Diseases 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/35—Clustering; Classification
- G06F16/353—Clustering; Classification into predefined classes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
- G06F18/253—Fusion techniques of extracted features
-
- 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
-
- 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
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/10—Services
- G06Q50/20—Education
- G06Q50/205—Education administration or guidance
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2216/00—Indexing scheme relating to additional aspects of information retrieval not explicitly covered by G06F16/00 and subgroups
- G06F2216/03—Data mining
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Evolutionary Computation (AREA)
- Molecular Biology (AREA)
- Educational Administration (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Tourism & Hospitality (AREA)
- Strategic Management (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Educational Technology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Databases & Information Systems (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Economics (AREA)
- Human Resources & Organizations (AREA)
- Marketing (AREA)
- Primary Health Care (AREA)
- General Business, Economics & Management (AREA)
- Electrically Operated Instructional Devices (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Machine Translation (AREA)
Abstract
The invention discloses a device and a method for analyzing text feedback fine granularity of students in combination with deep learning, which relate to the field of education, and the device comprises: a data acquisition module that acquires text data; the text embedding module is connected with the data acquisition module, performs word embedding processing on the text data and outputs word vectors; and the deep learning module is connected with the text embedding module, captures the multi-scale features of the word vectors by using a convolutional neural network, transmits the multi-scale features into a weighted bidirectional pyramid network for feature fusion, transmits the fused features into a text category classifier and a text polarity classifier, and respectively outputs category classification and emotion polarity of the text data. The invention adds category classification analysis on the basis of emotion polarity classification, thereby more fully utilizing the information in the feedback data.
Description
Technical Field
The invention relates to the field of education, in particular to a device and a method for analyzing text feedback fine granularity of students in combination with deep learning.
Background
Against the background that large shifts are the mainstream mode of education, it is clear contrast that there is limited teacher effort and time. The most effort is to take care of every student and know the latest learning condition of every student, and a great deal of information feedback and interaction are needed, but teachers are full of mind and insufficient in strength and almost impossible to achieve. Meanwhile, many schools have home-school communication originally to strengthen the communication between both education parties, and the online education platform also provides a discussion board of courses for students to communicate online. The online and online communication of both sides of education provides information support for better caring for each student, and the effective analysis of the communication information is beneficial to helping teachers to know the learning requirements and states of students more quickly and comprehensively, to take medicine according to symptoms and to take personalized teaching guidance measures.
Currently, many student feedback analyses for offline and online education are limited to polarity analysis of student evaluations, i.e., whether the student's evaluation results for the course are ultimately positive, negative or neutral. The results of the polarity analysis do not take full advantage of the information in the data and the additional value that the data produces.
Therefore, technical personnel in the field are dedicated to develop a student text feedback fine-grained analysis device and method integrating deep learning, information in data is more fully mined, teachers are helped to learn learning requirements and states of students more quickly and comprehensively, time of the teachers can be saved, teaching burden is relieved, the teachers can be assisted to provide students with refined learning guidance, and work efficiency of the teachers and learning efficiency of the students are improved.
Disclosure of Invention
In view of the problem that the text feedback of students is not fully utilized for analysis in the prior art and the value of the text feedback of the students is underestimated, the technical problem to be solved by the invention is how to fully analyze the text feedback data of the students, thereby assisting teachers in teaching work, fully knowing the learning conditions of the students and providing targeted guidance.
In order to achieve the purpose, the invention provides a student text feedback fine-grained analysis device and method integrating deep learning, which are used for completing the analysis of student feedback and assisting teachers in teaching work, so that students can fully know the learning conditions and provide targeted guidance.
The invention provides a student text feedback fine-grained analysis device integrating deep learning, which comprises:
a data acquisition module that acquires text data;
the text embedding module is connected with the data acquisition module, performs word embedding processing on the text data and outputs word vectors;
and the deep learning module is connected with the text embedding module, captures the multi-scale features of the word vectors by using a convolutional neural network, transmits the multi-scale features into a weighted bidirectional pyramid network for feature fusion, transmits the fused features into a text category classifier and a text polarity classifier, and respectively outputs category classification and emotion polarity of the text data.
Further, the word embedding processing of the text embedding module is to obtain a word vector by training a continuous bag-of-words model, add position information, encode the position of each character in the text data, finally fuse part-of-speech tag embedding, and combine the word vector obtained by training the continuous bag-of-words model and a position-encoded vector to obtain a final word vector.
Further, the input of the continuous bag of words model is a word vector corresponding to a context-dependent word of the current central word, and the output of the continuous bag of words model is the word vector of the current central word.
Further, the data acquisition module identifies text information communicated offline by using an optical character identification scanner and stores the text information as the text data, or crawls text information communicated online by using crawler codes and stores the crawled text information as the text data.
Further, still include:
the data preprocessing module is arranged between the data acquisition module and the text embedding module, directly discards data with incomplete information in the text data, deletes punctuation marks and stop words in the text data, and only retains pure character information.
Further, still include:
and the post-processing module is arranged behind the deep learning module and is used for storing the text data, the category classification and the emotion polarity.
Further, the text data, the category classification and the emotion polarity are stored in the form of a three-dimensional tuple < T, C, P >, wherein: t represents a text, C represents the category classification of the evaluation content in the text, and P represents the emotion polarity embodied in the feedback.
The invention provides a student text feedback fine-grained analysis method integrating deep learning, which comprises the following steps:
step 1, acquiring text data by adopting different means according to different online and offline scenes;
step 2, performing word embedding processing on the text data, training by using a continuous bag-of-words model to obtain a word vector, adding position information, coding the position of each character in the text data, finally embedding a part-of-speech tag, and combining the word vector obtained by training the continuous bag-of-words model and a position-coded vector to obtain a final word vector;
and 3, capturing multi-scale features of the word vector by using a convolutional neural network, transmitting the multi-scale features into a weighted bidirectional pyramid network for feature fusion, transmitting the fused features into a text category classifier and a text polarity classifier, and respectively outputting category classification and emotion polarity of the text data.
Further, a support vector machine is used in said step 3 to train said text category classifier and said text polarity classifier.
Further, in the step 3, according to different characteristics of the offline and the online, the category of the offline is divided into: teaching methods, teaching attitudes and self-reflexive provinces, and the categories of the online categories are divided into: teaching method, teaching attitude, self-reflexive province and platform use sense; the polarity classes are divided into: positive, neutral and negative; and respectively comparing the prediction results of the text category classifier and the text polarity classifier with the label data, calculating a cross entropy loss function, and updating the weight in the network model by using a random gradient descent method.
The invention provides a device and a method for analyzing the text feedback fine granularity of a student integrating deep learning, which at least have the following technical effects:
1. in the past, the processing of the feedback of students is only limited to the evaluation polarity classification, and the invention adds the category classification analysis on the basis of the emotion polarity classification, thereby more fully utilizing the information in the feedback data;
2. the method has the advantages that the text feedback of the students is analyzed in a finer granularity, the results are output and stored for reference of teachers, so that the teachers can conveniently adjust teaching methods for individual students in time, feedback information of the students can be fully utilized, the students can more possibly obtain more remarkable score improvement under personalized guidance, the analysis results can be stored, the effect comparison of the teachers after corresponding measures are taken by utilizing the analysis results can be tracked, and extra value generated by data is better achieved.
The conception, the specific structure and the technical effects of the present invention will be further described with reference to the accompanying drawings to fully understand the objects, the features and the effects of the present invention.
Drawings
FIG. 1 is a flow chart of a preferred embodiment of the present invention;
fig. 2 is a block diagram of the deep learning module of the embodiment shown in fig. 1.
Detailed Description
The technical contents of the preferred embodiments of the present invention will be more clearly and easily understood by referring to the drawings attached to the specification. The present invention may be embodied in many different forms of embodiments and the scope of the invention is not limited to the embodiments set forth herein.
Currently, many student feedback analyses for offline and online education are limited to polarity analysis of student evaluations, i.e., whether the student's evaluation results for the course are ultimately positive, negative or neutral. The results of the polarity analysis do not take full advantage of the information in the data and the additional value that the data produces.
Therefore, technical personnel in the field are dedicated to develop a student text feedback fine-grained analysis device and method integrating deep learning, information in data is more fully mined, teachers are helped to learn learning requirements and states of students more quickly and comprehensively, time of the teachers can be saved, teaching burden is relieved, the teachers can be assisted to provide students with refined learning guidance, and work efficiency of the teachers and learning efficiency of the students are improved. Specifically, the device and the method provided by the invention are combined with a supervised learning method, and a Convolutional Neural Network (CNN) is used for capturing multi-scale features of a text, and then the features with different scales are transmitted into a Weighted Bi-directional Feature Pyramid Network (BiFPN) for Feature fusion. And one of the input text category classifiers is used for classifying the category of the output text, and the other input text polarity classifier is used for outputting the emotion polarity of the text.
In order to achieve the above object, the present invention provides a device for analyzing fine-grained feedback of a student text fused with deep learning, comprising:
the data acquisition module acquires text data;
the text embedding module is connected with the data acquisition module, performs word embedding processing on the text data and outputs word vectors;
and the deep learning module is connected with the text embedding module, captures multi-scale features of word vectors by using a convolutional neural network, transmits the multi-scale features into a weighted bidirectional pyramid network for feature fusion, transmits the fused features into a text category classifier and a text polarity classifier, and respectively outputs category classification and emotion polarity of text data.
The word embedding processing of the text embedding module is to train a Continuous Bag-of-Words (CBOW) model to obtain a word vector, add position information, encode the position of each character in text data, and finally fuse part-of-speech tag embedding, so that the following neural network can more fully obtain semantic information of a learning word, and combine the semantic information with the word vector obtained by training the Continuous Bag-of-Words model and the position-encoded vector to form a final word vector. The input of the continuous bag-of-words model is a word vector corresponding to the context-dependent word of the current central word, and the output of the continuous bag-of-words model is the word vector of the current central word.
The data acquisition module is a module for acquiring data by the whole system, and for text data of an offline home-school communication book, an Optical Character Recognition scanner (OCR) is used for recognizing text information and storing the information into the system; for an online education platform on line, a crawler code written by Python is used for crawling and storing the student feedback information of the website platform to be analyzed into the system.
The invention provides a student text feedback fine-grained analysis device integrating deep learning, which further comprises: and a data preprocessing module.
The data preprocessing module is arranged between the data acquisition module and the text embedding module, directly discards data with incomplete information in the text data, deletes punctuation marks and stop words in the text data, and only retains pure character information. Specifically, the data acquired by the data acquisition module is noisy and requires further cleaning. The data preprocessing module directly discards data with incomplete information; second, punctuation marks in the data as well as stop words such as "in", "also", etc. are deleted, leaving only pure character information.
The invention provides a student text feedback fine-grained analysis device integrating deep learning, which further comprises: and a post-processing module.
The post-processing module is arranged behind the deep learning module and used for storing text data, category classification and emotion polarity. Wherein the text data, category classification and emotion polarity are stored in the form of a three-dimensional tuple < T, C, P >, wherein: t represents a text, C represents category classification of evaluation content in the text, and P represents emotion polarity embodied in feedback.
The invention provides a student text feedback fine-grained analysis method integrating deep learning, which comprises the following steps:
firstly, obtaining student feedback text data to be processed, then carrying out basic preprocessing on the data, removing noise information and punctuation marks in original data, stopping words and the like, obtaining cleaned data, converting the data into word vectors capable of being processed by a computer in a word embedding mode, transmitting the word vectors into a deep learning model, extracting features of the text by the model, then using the extracted features as the input of two classifiers, enabling the two classifiers to respectively output the category to which the information in the text belongs and the emotional tendency of students reflected in the information, and finally storing the predicted classification result into a special data analysis completion folder.
The invention provides a specific embodiment of a student text feedback fine-grained analysis method integrating deep learning, which comprises the following steps:
step 1, acquiring text data by adopting different means according to different online and offline scenes;
step 2, performing word embedding processing on the text data, utilizing a continuous bag-of-words model to train to obtain a word vector, simultaneously adding position information, coding the position of each character in the text data, finally fusing part-of-speech tag embedding, and combining the word vector obtained by the training of the continuous bag-of-words model and the position-coded vector to form a final word vector;
and 3, capturing multi-scale features of the word vectors by using a convolutional neural network, transmitting the multi-scale features into a weighted bidirectional pyramid network for feature fusion, transmitting the fused features into a text category classifier and a text polarity classifier, and respectively outputting category classification and emotion polarity of text data.
Wherein a support vector machine is used in step 3 to train the text category classifier and the text polarity classifier.
In step 3, according to different characteristics of the offline and the online, the category of the offline is divided into: teaching methods, teaching attitudes and self-reflexive provinces, and on-line categories are divided into: teaching method, teaching attitude, self-reflexive province and platform use sense; the polarity classes are divided into: positive, neutral and negative; and respectively comparing the prediction results of the text category classifier and the text polarity classifier with the label data, calculating a cross entropy loss function, and updating the weight in the network model by using a random gradient descent method.
In another specific embodiment of the method for analyzing the text feedback fine granularity of the student with the deep learning fused (as shown in fig. 1), the method comprises the following steps:
step 1, acquiring data to be analyzed by different means according to different online and offline scenes, writing codes by python for capturing the online data, and acquiring the offline data by using the conventional optical character recognition scanner;
step 2, after the system obtains the data, cleaning the data, deleting the data which do not meet the requirements and redundant punctuation marks, stop words and the like in each data, and removing noise interference in the unstructured data as much as possible;
step 3, the text embedding module receives the preprocessed data, converts the data into a word vector form, adds the position information and the part-of-speech tag information of each word in the text, and inputs the position information and the part-of-speech tag information into a subsequent module for further processing and analysis;
step 4, the deep learning module is a core module of the system, and the model is obtained through supervised learning training (as shown in fig. 2);
step 401, extracting features of different scales from data represented by word vectors by using a convolutional neural network as a feature extractor, and fusing the features of different scales through a weighted bidirectional feature pyramid network to improve network efficiency and input the fused features as common features of a subsequent classifier;
step 402, training two different classifiers by using a machine learning algorithm such as a support vector machine and the like, wherein one classifier classifies the category of an evaluation object in data, contends for different characteristics below and above the line, and divides the category below the line into a teaching method (evaluation on the teaching method related aspect of a teacher), a teaching attitude (evaluation on the teaching attitude related aspect of the teacher) and self-reflexion (self-reflexive evaluation of students in the learning process); because the online education is generally built on a platform, the use sense category of the platform is required to be added to classify the use senses of students on the platform during online learning; the other classifier is used for predicting the evaluation polarity of the evaluation object in the data and is roughly divided into three types of polarities, namely positive, neutral and negative. For example, there are student text ratings: the teacher's pace of class is too fast, i can not follow up a little and can not digest. "then from" teacher's rhythm of class too fast "in this text can predict that the category classification of the evaluation object belongs to the teaching method, from" i can't catch up with something, can't digest "can predict that the polarity of the evaluation object belongs to the negative evaluation. In the teaching mode, the prediction results of the two classifiers are respectively compared with the label data, a cross entropy loss function is calculated, and then a random gradient descent method is applied to update the weight in the network;
step 403, adjusting the network weight to a proper value in the training phase;
in actual use, only the word vector generated in the step 3 is needed to be input, and then two needed classification prediction results can be output;
and 5, the post-processing module is responsible for combining the results predicted by the two classifiers and the corresponding evaluation entity object into a ternary tuple form < T, C and P > to be stored locally, and a user can customize the name of the folder.
Compared with the prior art, the invention has the beneficial effects that:
the prior processing of student feedback including evaluation is limited to the polarity classification of evaluation, but no way is available for directly utilizing the polarity analysis results to make corresponding measures to improve teaching, and other more interactive data needs to be further analyzed. Therefore, the embodiment of the invention provides a method for analyzing the fine granularity of the text feedback of students in combination with deep learning, the fine granularity of the text feedback of the students is analyzed, and the result is output and stored for reference of teachers, so that the teaching method for individual students and the like can be adjusted in time, the feedback information is fully utilized, the students can obtain more remarkable grade promotion under personalized guidance, and the effect comparison after the students use of the analysis results by teachers to take corresponding measures can be tracked after the feedback information is stored in a system. In addition, the embodiment of the invention prominently solves the problems that the text feedback data of the students are not thoroughly analyzed and cannot be directly utilized by teachers to formulate and adjust learning schemes for the students, establishes a more perfect analysis result, can assist the teachers to better understand the students and indirectly help the achievements of the students to be improved.
The foregoing detailed description of the preferred embodiments of the invention has been presented. It should be understood that numerous modifications and variations could be devised by those skilled in the art in light of the present teachings without departing from the inventive concepts. Therefore, the technical solutions available to those skilled in the art through logic analysis, reasoning and limited experiments based on the prior art according to the concept of the present invention should be within the scope of protection defined by the claims.
Claims (10)
1. The utility model provides a student's text feedback fine grit analytical equipment who fuses deep learning which characterized in that includes:
a data acquisition module that acquires text data;
the text embedding module is connected with the data acquisition module, performs word embedding processing on the text data and outputs word vectors;
and the deep learning module is connected with the text embedding module, captures the multi-scale features of the word vectors by using a convolutional neural network, transmits the multi-scale features into a weighted bidirectional pyramid network for feature fusion, transmits the fused features into a text category classifier and a text polarity classifier, and respectively outputs category classification and emotion polarity of the text data.
2. The student text feedback fine-grained analysis device integrating deep learning as claimed in claim 1, wherein the word embedding process of the text embedding module is to obtain a word vector by training a continuous bag-of-words model, add position information, encode the position of each character in the text data, finally embed a word-property label, and combine the word vector obtained by training the continuous bag-of-words model and a position-encoded vector to form a final word vector.
3. The apparatus for feedback fine-grained analysis of student text fused with deep learning according to claim 2, wherein the input of the continuous bag of words model is a word vector corresponding to a context-dependent word of a current central word, and the output of the continuous bag of words model is the word vector of the current central word.
4. The student text feedback fine-grained analysis device integrating deep learning as claimed in claim 1, wherein the data acquisition module identifies text information communicated offline by using an optical character recognition scanner and stores the text information as the text data, or crawls text information communicated online by using crawler codes and stores the text information as the text data.
5. The apparatus for student text feedback fine-grained analysis with deep learning fused as claimed in claim 1, further comprising:
the data preprocessing module is arranged between the data acquisition module and the text embedding module, directly discards data with incomplete information in the text data, deletes punctuation marks and stop words in the text data, and only retains pure character information.
6. The apparatus for student text feedback fine-grained analysis with deep learning fused as claimed in claim 1, further comprising:
and the post-processing module is arranged behind the deep learning module and is used for storing the text data, the category classification and the emotion polarity.
7. The student text feedback fine-grained analysis device with deep learning fused as claimed in claim 6 wherein the text data, the category classification and the emotion polarity are stored in the form of a three-dimensional tuple < T, C, P >, wherein: t represents a text, C represents the category classification of the evaluation content in the text, and P represents the emotion polarity embodied in the feedback.
8. A student text feedback fine-grained analysis method integrating deep learning is characterized by comprising the following steps:
step 1, acquiring text data by adopting different means according to different online and offline scenes;
step 2, performing word embedding processing on the text data, training by using a continuous bag-of-words model to obtain a word vector, adding position information, coding the position of each character in the text data, finally embedding a part-of-speech tag, and combining the word vector obtained by training the continuous bag-of-words model and a position-coded vector to obtain a final word vector;
and 3, capturing multi-scale features of the word vector by using a convolutional neural network, transmitting the multi-scale features into a weighted bidirectional pyramid network for feature fusion, transmitting the fused features into a text category classifier and a text polarity classifier, and respectively outputting category classification and emotion polarity of the text data.
9. The method for student text feedback fine-grained analysis with deep learning fused as claimed in claim 8, wherein a support vector machine is used in step 3 to train the text category classifier and the text polarity classifier.
10. The method for analyzing the text feedback fine granularity of the student fusing the deep learning according to the claim 8, wherein in the step 3, the category under the line is classified into the following categories according to different characteristics under the line and on the line: teaching methods, teaching attitudes and self-reflexive provinces, and the categories of the online categories are divided into: teaching method, teaching attitude, self-reflexive province and platform use sense; the polarity classes are divided into: positive, neutral and negative; and respectively comparing the prediction results of the text category classifier and the text polarity classifier with the label data, calculating a cross entropy loss function, and updating the weight in the network model by using a random gradient descent method.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110059701.2A CN112765351A (en) | 2021-01-18 | 2021-01-18 | Deep learning-fused student text feedback fine-grained analysis device and method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110059701.2A CN112765351A (en) | 2021-01-18 | 2021-01-18 | Deep learning-fused student text feedback fine-grained analysis device and method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112765351A true CN112765351A (en) | 2021-05-07 |
Family
ID=75702353
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110059701.2A Pending CN112765351A (en) | 2021-01-18 | 2021-01-18 | Deep learning-fused student text feedback fine-grained analysis device and method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112765351A (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111353042A (en) * | 2020-02-27 | 2020-06-30 | 浙江大学 | Fine-grained text viewpoint analysis method based on deep multi-task learning |
| CN111680154A (en) * | 2020-04-13 | 2020-09-18 | 华东师范大学 | Comment text attribute level emotion analysis method based on deep learning |
-
2021
- 2021-01-18 CN CN202110059701.2A patent/CN112765351A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111353042A (en) * | 2020-02-27 | 2020-06-30 | 浙江大学 | Fine-grained text viewpoint analysis method based on deep multi-task learning |
| CN111680154A (en) * | 2020-04-13 | 2020-09-18 | 华东师范大学 | Comment text attribute level emotion analysis method based on deep learning |
Non-Patent Citations (1)
| Title |
|---|
| 王会岛: "基于主题的学生评教短文本多粒度情感分析", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111026842B (en) | Natural language processing method, natural language processing device and intelligent question-answering system | |
| CN112085012B (en) | Project name and category identification method and device | |
| CN106886580B (en) | Image emotion polarity analysis method based on deep learning | |
| CN110781680A (en) | Semantic similarity matching method based on twin network and multi-head attention mechanism | |
| CN108664932B (en) | Learning emotional state identification method based on multi-source information fusion | |
| CN111581361A (en) | Intention identification method and device | |
| CN107833603A (en) | Electronic medical record document sorting technique, device, electronic equipment and storage medium | |
| CN112182166A (en) | Text matching method and device, electronic equipment and storage medium | |
| CN108763384A (en) | For the data processing method of text classification, data processing equipment and electronic equipment | |
| CN112749562A (en) | Named entity identification method, device, storage medium and electronic equipment | |
| CN106599824B (en) | A kind of GIF animation emotion identification method based on emotion pair | |
| CN112988963A (en) | User intention prediction method, device, equipment and medium based on multi-process node | |
| CN112905739A (en) | False comment detection model training method, detection method and electronic equipment | |
| CN110705490A (en) | Visual emotion recognition method | |
| Rao et al. | A first look: Towards explainable textvqa models via visual and textual explanations | |
| CN115526166A (en) | Image-text emotion inference method, system, storage medium and equipment | |
| CN113312912A (en) | Machine reading understanding method for traffic infrastructure detection text | |
| Leong | Deep learning of facial embeddings and facial landmark points for the detection of academic emotions | |
| CN114417785A (en) | Knowledge point annotation method, model training method, computer device, and storage medium | |
| CN111382243A (en) | Text category matching method, text category matching device and terminal | |
| CN116758558A (en) | Cross-modal generation countermeasure network-based image-text emotion classification method and system | |
| CN116741396A (en) | Article classification method and device, electronic equipment and storage medium | |
| CN112765351A (en) | Deep learning-fused student text feedback fine-grained analysis device and method | |
| CN114974571A (en) | Multimode data-driven autism detection system, device and storage medium | |
| Abdulhamied et al. | Real-time recognition of American sign language using long-short term memory neural network and hand detection |
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: 20210507 |