CN107918487A - A kind of method that Chinese emotion word is identified based on skin electrical signal - Google Patents
A kind of method that Chinese emotion word is identified based on skin electrical signal Download PDFInfo
- Publication number
- CN107918487A CN107918487A CN201710998545.XA CN201710998545A CN107918487A CN 107918487 A CN107918487 A CN 107918487A CN 201710998545 A CN201710998545 A CN 201710998545A CN 107918487 A CN107918487 A CN 107918487A
- Authority
- CN
- China
- Prior art keywords
- skin
- chinese
- emotion
- value
- algorithm
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
- G06F3/01—Input arrangements or combined input and output arrangements for interaction between user and computer
- G06F3/011—Arrangements for interaction with the human body, e.g. for user immersion in virtual reality
- G06F3/015—Input arrangements based on nervous system activity detection, e.g. brain waves [EEG] detection, electromyograms [EMG] detection, electrodermal response detection
-
- 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
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/02—Preprocessing
- G06F2218/04—Denoising
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/08—Feature extraction
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/12—Classification; Matching
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Data Mining & Analysis (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Neurosurgery (AREA)
- Biophysics (AREA)
- Neurology (AREA)
- Databases & Information Systems (AREA)
- Human Computer Interaction (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Dermatology (AREA)
- Computational Linguistics (AREA)
- Evolutionary Computation (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)
Abstract
The invention discloses a kind of method that Chinese emotion word is identified based on skin electrical signal.The advantage of physiological parameter identification emotion is used to identify Chinese emotion word by this method.Specifically gather comprising skin pricktest, the data after collection are pre-processed, feature extraction, normalized, feature selecting, using improved simulated annealing artificial neural network algorithm obtaining classification results, emotion word is finally added in classification results to compare, and is identified.As embodiment, the present invention be based on from《Modern Chinese dictionary》、《Modern Chinese classified dictionary》、《New century Conforming of Chinese New Words dictionary》In 50 highest emotion words of emotion intensity filtering out identified.Experiment proves that the present invention can complete the identification to Chinese emotion word and accuracy is very high, fully show that using extraction of the physiological parameter to text emotion word be feasible, new thinking is provided for later stage text analyzing, and present system framework is clear, simple, it is easy to accomplish.
Description
Technical Field
The invention belongs to the field of processing natural language data by an electrical digital data processing technology, and particularly relates to a method for recognizing Chinese emotion words based on a skin electrical signal emotion recognition technology.
Background
The use of skin electrical signals to identify emotion has its unique advantages, such as being physiological parameters and therefore more objective, such as being easier to collect than other physiological parameters, being most effective and sensitive to neuro-emotional changes. For studying emotions by using skin electrical signals, the current technology is relatively mature, so that the technology is timely used for introducing text analysis. The current text analysis has the defects of strong subjectivity, difficult syntax and semantic splitting, imperfect emotion word stock and the like, which hinder the pace of text emotion analysis.
However, with the vigorous development of the internet, the text information is very expensive, and contains a great amount of useful and practical information, so people are reluctant to abandon the extraction of the text information. The Chinese emotional words need to be objectively identified by using more objective physiological parameters, which inevitably provides a new idea for simple text analysis.
Disclosure of Invention
The invention aims to provide a method for identifying Chinese emotion words by using skin electric signals, which provides a brand-new idea for simple text emotion analysis and enables the text emotion analysis to be more accurate.
In order to achieve the purpose, the technical scheme adopted by the invention is a method for identifying Chinese emotional words based on skin electric signals, which specifically comprises the following steps:
s1: collecting skin electricity;
s2: preprocessing the acquired data;
s3: extracting characteristics;
s4: normalization processing;
s5: selecting characteristics;
s6: obtaining a classification result by utilizing an improved simulated annealing artificial neural network algorithm;
s7: and adding emotion word comparison in the classification result for identification.
Preferably, the preprocessing in step 2 is performed by denoising using wavelet transform.
Further, in the feature extraction in step 3, statistical values which can represent changes of skin electrical signals in time domain and frequency domain of the signals are extracted as original features of emotion recognition research.
Further, the time domain original features include a mean value, a median value, a maximum value, a minimum value, a standard deviation, a minimum value ratio, a maximum and minimum difference value of the skin electric signal, and 24 time domain features generated by respectively performing first-order difference and second-order difference calculation on the signal features and then extracting the statistical features.
Further, before the frequency domain features are extracted, discrete Fourier transform is carried out on the skin electric signals, and then the mean value, the median value, the standard deviation, the maximum value, the minimum value and the maximum and minimum difference value of the frequency are calculated to obtain 6 frequency domain features.
Further, the normalization processing in the step 4 makes the value range of each characteristic value limited between 0 and 1, and the method for removing the individual differences is as follows:
wherein X G In order to be the original signal, the signal is,for the mean value of each subject at rest, normalization was followed to yield:
X＝(X G -X mean )/(X max -X min ) (2)。
further, in order to identify the emotion with the least number of features and the highest identification rate, in the step 5, a plurality of groups are randomly selected from the normalized data during feature selection and are divided into three parts, wherein the first part is a classifier training set, the second part is a testing set and is used for testing the classification effect, and the last part of data is used for verifying the effectiveness of the feature set in emotion identification.
Further, the improved simulated annealing artificial neural network algorithm comprises the following steps:
the method comprises the following steps: determining a neural network structure from the input and output of the sample;
step two: a simulated annealing algorithm with memory is applied, and the method specifically comprises the following steps:
1) Parameters are initialized, thus generating initial weight S 0 At this time, the initial temperature T is set 0 >, 0, iteration number i =0, checking precision epsilon, and making f out ＝f(S 0 )，f * ＝f(S 0 ),S p ＝S 0 ；
2) Weighting S of network p As an initial starting point S 0 Optimizing according to Powell algorithm, and quickly searching a certain local minimum value point;
3) Setting memory variables x 'and f (x') for memorizing the optimal solution and the optimal objective function value encountered currently, respectively, and initializing x 'and f (x') to be equal to the initial solution x when the algorithm is just started 0 And its objective function value f (x) 0 ) After the iteration starts, the objective function value f (x) is used each time a new search solution is accepted k ) Comparing with f (x'), if f (x) k ) Better than f (x'), then x is used respectively k And f (x) k ) Replacing the original x 'and f (x'), and finally obtaining a global optimal solution when the algorithm is finished;
4) The new group of network weights S is obtained p Let S i ＝S p ,f out ＝f(S i )，f * ＝f(S i ) The network weight S i As an iteration value x, let the current solution S i = x, let T = T i Annealing to obtain a new set of network weights S i+1 According to T i ＝T 0 /(1 + ln (i)) annealed, i = i +1;
5) If the requirement or the number of iterations is met after annealing, the algorithm ends, if f (S) i )<f out Let S stand out p ＝S i+1 Returning to the step 4;
training and predicting a neural network, wherein the training is to determine a network structure by setting fixed input and output, the neural network continuously adjusts the connection weight between each neuron in the training process so as to reduce the error between the training output and the designated output, and the prediction is the process of processing input data by the trained network to obtain output;
step four: and finally, comparing the output result with the form information input by the testee in the experimental process to finish the identification and comparison of the emotional words.
Compared with the prior art, the invention has the beneficial effects that:
1, the method can complete the recognition of the Chinese emotional words with high accuracy, and basically achieves the expected result.
2, the method fully shows that the extraction of the text sentiment words by using the physiological parameters is feasible, and provides a new idea for the later text analysis.
3, the system of the invention has clear and simple structure and is easy to realize.
Drawings
Figure 1 shows a schematic flow diagram of the overall scheme.
FIG. 2 shows a partial emotion vocabulary questionnaire.
Fig. 3 shows an experimental emotion word recognition table.
FIG. 4 shows the alignment chart of the two identifications.
Detailed Description
The present invention will now be described in further detail with reference to the accompanying drawings and examples.
As an embodiment, the method firstly screens 2000 emotional words from a modern Chinese dictionary, a modern Chinese classification dictionary and a new Chinese word dictionary, and then screens 100 most frequently used emotional words from the 2000 emotional words. And finally, screening the 100 words again to obtain 50 emotional words with the highest emotional intensity.
20 people in a certain laboratory are collected by using a laboratory skin electricity collecting tool. The laboratory has 20 people all have good health, no history of heart disease mental illness, no psychotropic drugs are taken within one year, and all people in all ages from 20 to 50 years old. The experimental materials are 50 selected emotional words with strong emotions, the tested person is required to sit in front of the computer screen, one emotional word appears on the computer screen every 40 seconds at the moment, and the tested person is required to associate scenes related to the emotional words when the emotional words appear. The first 30 seconds are used to associate and the last 10 seconds are used to fill in whether there is a sensation and fill in the emotional intensity (0 strong, 1 strong, 2 normal, 3 weak, 4 weak). And then played in sequence until the playing of 50 emotional words is completed.
The collected data is preprocessed, and the collected skin electric signals need to be denoised because the skin electric signals are weak and are easily influenced by machine interference, myoelectric interference, electromagnetic interference and the like. The invention adopts wavelet transformation to carry out denoising treatment. The wavelet transform has higher frequency resolution and lower time resolution in the low frequency part and higher time resolution and lower frequency resolution in the high frequency part, so that the wavelet transform has adaptivity to signals and is very suitable for analysis of physiological signals.
Before the experiment starts, the flow and purpose of the experiment are explained in detail to the testee. First, each subject is required to sit 80cm directly in front of the computer display screen. The experiment was started. The testee is required to close eyes for one minute and then open the eyes to see the screen, and at the moment, one emotional word appears on the screen every 40 seconds, wherein the emotional word is displayed for 30 seconds, and the emotional word emotion intensity table is filled in for 10 seconds. When the emotional words appear, the testee associates with the corresponding scene under the stimulation of the emotional words, fills the questionnaire when the screen is blank, and leads the emotion to be calm. And playing the selected 50 emotional words in sequence until the end.
After deleting the invalid data, 270 sets of valid data are screened out. With reference to the feature extraction method at the university of Augsburg in Germany, statistical values which are most representative of the change of skin electrical signals in the time domain and the frequency domain of the signals are extracted as the original features of emotion recognition research. In the time domain, 22 time domain features such as the maximum value, the minimum value, the standard deviation, the first-order difference minimum value ratio, the second-order difference standard deviation, the second-order difference minimum value ratio and the like of the skin electric signal are extracted. In order to extract the frequency domain features of the skin electrical signals, discrete Fourier transform is performed on the skin electrical signals, and then frequency mean values, median values, standard deviations, maximum values, minimum values and maximum and minimum difference values are calculated to obtain 6 frequency domain features.
Because the skin electric signals have large individual difference, and the value ranges of the characteristic values of all the statistical characteristics extracted according to the formula are in different orders of magnitude, the statistical distribution of data is normalized and the subsequent processing is facilitated in order to facilitate uniform comparison and normalization processing of all the characteristics, so that the value ranges of all the characteristic values are limited between 0 and 1. The formula is as follows:
method for removing individual differences:
wherein X G In the form of an original signal, the signal,is each timeMean value of the resting state of individual subjects.
After normalization, the following results are obtained:
X＝(X G -X mean )/(X max -X min ) (2)
and performing feature selection on the processed data. The feature selection is to identify the emotion with the least number of features and the highest recognition rate. Randomly selecting 180 groups from the 270 groups of normalized data, and dividing the data into three parts, namely forming a classifier training set by the first 80 groups of data; the middle 60 groups of data form a test set to test the classification effect; the last 40 sets of data were used to verify the validity of the feature set in emotion recognition.
A neural network algorithm is a distributed storage of information, which is stored throughout the network, where not one external information, but part of the contents of multiple information, is stored somewhere on the network. The advantage of the neural network algorithm is obvious, because the information is stored on the network, the information storage and processing are combined into one, so that the data can be massively processed in parallel, and the method has strong fault-tolerant capability and robustness. It also has autodidactic and adaptive properties. However, the neural network algorithm has a significant drawback that it is easy to fall into a local minimum and the convergence process is slow, so that improvement is required.
Because of the defects of the neural network algorithm, the function of the algorithm is greatly improved by adding the simulated annealing technology, and the defects are solved to a great extent. The method firstly uses Powell algorithm to quickly converge to the local minimum value, when the local minimum value is found, the simulated annealing search strategy is used to immediately search whether the local minimum value has a valley bottom or not, the searching is carried out for a plurality of times, at the moment, the memorability is added, the searched optimal value and the optimal function value are recorded, and thus, the global optimal value can be quickly found. The method is applied to a neural network algorithm, so that the defect that the neural network can not jump out of the local minimum value is well overcome.
The improved simulated annealing artificial neural network algorithm steps are summarized as follows:
determining a neural network structure according to the input and the output of the sample.
Step two: a simulated annealing algorithm with memory is applied:
1) And initializing parameters. This results in an initial weight S 0 At this time, the initial temperature T is set 0 >, 0, iteration times i =0, and checking precision epsilon. Let f be out ＝f(S 0 )，f * ＝f(S 0 ),S p ＝S 0 。
2) Weighting S of network p As an initial starting point S 0 And optimizing according to a Powell algorithm, and quickly searching a certain local minimum value point.
3) And setting memory variables x 'and f (x') for memorizing the optimal solution and the optimal objective function value which are met currently respectively. The algorithm initially initializes x 'and f (x') to be equal to the initial solution x, respectively 0 And its objective function value f (x) 0 ) After the iteration starts, the objective function value f (x) is used each time a new search solution is accepted k ) Comparing with f (x'), if f (x) k ) Better than f (x'), then x is used respectively k And f (x) k ) Instead of the original x 'and f (x'). And finally, obtaining a global optimal solution when the algorithm is finished.
4) The new group of network weights S is obtained p Let S stand out i ＝S p ,f out ＝f(S i )，f * ＝f(S i ). Weighting S of network i As the iteration value x, let the current solution S i = x, let T = T i And carrying out annealing operation. Obtain a new set of network weights S i+1 . According to T i ＝T 0 /(1 +ln (i)) is annealed. i = i +1.
5) If the requirements or iteration times are met after annealing, the algorithm ends. If f (S) i )<f out Let S p ＝S i+1 And (4) is turned back.
And step three, training and predicting the neural network. Training is to determine the network structure by setting fixed inputs and outputs. In the training process, the neural network continuously adjusts the connection weight between each neuron so as to reduce the error between the training output and the designated output. Prediction is the process of processing input data by a trained network to obtain output.
And finally, comparing the output result with a table filled by a testee in the experimental process to finish the identification and comparison of the emotional words. Wherein, part of the emotion word questionnaire is shown in fig. 2, the emotion word recognition graph in the experiment is shown in fig. 3, and the comparison recognition graph of the two is shown in fig. 4. The experimental result shows that the identification of the Chinese emotional words can be basically completed by utilizing the change of the skin electric signal.
Claims (8)
1. A method for recognizing Chinese emotional words based on skin electric signals is characterized by comprising the following steps:
s1: collecting skin electricity;
s2: preprocessing the acquired data;
s3: extracting characteristics;
s4: normalization processing;
s5: selecting characteristics;
s6: obtaining a classification result by utilizing an improved simulated annealing artificial neural network algorithm;
s7: and adding emotion word comparison in the classification result for identification.
2. The method for recognizing Chinese emotional words based on the electrical skin signals as claimed in claim 1, wherein the preprocessing in step 2 is performed with a de-noising processing using wavelet transform.
3. The method for recognizing Chinese emotional words based on electrical skin signals as claimed in claim 1, wherein the feature extraction in step 3 is to extract statistics representing the change of electrical skin signals in time domain and frequency domain of the signals as the original features of the emotion recognition research.
4. The method for recognizing the Chinese emotional words based on the skin electric signals according to claim 3, wherein the time domain original features include a mean value, a median value, a maximum value, a minimum value, a standard deviation, a minimum value ratio, a maximum and minimum difference value of the skin electric signals, and 24 time domain features generated by respectively performing first-order difference calculation and second-order difference calculation on the signal features and then extracting the statistical features.
5. The method for identifying Chinese emotional words based on the skin electrical signals as claimed in claim 3, wherein before the frequency domain features are extracted, discrete Fourier transform is performed on the skin electrical signals, and then frequency mean, median, standard deviation, maximum, minimum, and maximum-minimum difference are calculated to obtain 6 frequency domain features.
6. The method for identifying Chinese emotion words based on skin electric signals as claimed in claim 1, wherein the normalization process in step 4 limits the value range of each eigenvalue to 0 to 1, and the method for removing individual differences is as follows:
wherein X G In order to be the original signal, the signal is,for the mean value of each subject at rest, normalization was followed to yield:
X＝(X G -X mean )/(X max -X min ) (2)。
7. the method for recognizing Chinese emotional words based on electrical skin signals as claimed in claim 1, wherein in the step 5, a plurality of groups are randomly selected from the normalized data and divided into three parts, wherein the first part is a classifier training set, the second part is a test set for testing the classification effect, and the last part of data is used for verifying the effectiveness of the feature set in the emotion recognition.
8. The method for recognizing Chinese emotion words based on skin electric signals as claimed in claim 1, wherein said modified simulated annealing artificial neural network algorithm comprises the steps of:
the method comprises the following steps: determining a neural network structure from the input and output of the sample;
step two: a simulated annealing algorithm with memory is applied, and the method specifically comprises the following steps:
1) Parameters are initialized, thus generating initial weight S 0 At this time, the initial temperature T is set 0 >, 0, iteration number i =0, checking precision epsilon, and making f out ＝f(S 0 )，f * ＝f(S 0 ),S p ＝S 0 ；
2) Weighting S of network p As an initial starting point S 0 Optimizing according to Powell algorithm, and quickly searching a certain local minimum value point;
3) Setting memory variables x 'and f (x') for memorizing the optimal solution and the optimal objective function value encountered currently, respectively, and initializing x 'and f (x') to be equal to the initial solution x when the algorithm is just started 0 And its objective function value f (x) 0 ) After the iteration starts, each time a new search solution is accepted, its objective function value f (x) is added k ) Comparing with f (x'), if f (x) k ) Better than f (x'), then x is used respectively k And f (x) k ) Replacing the original x 'and f (x'), and finally obtaining a global optimal solution when the algorithm is finished;
4) The new group of network weights S is obtained p Let S stand out i ＝S p ,f out ＝f(S i )，f * ＝f(S i ) The network weight S i As the iteration value x, let the current solution S i = x, let T = T i Annealing to obtain a new set of network weights S i+1 According to T i ＝T 0 /(1 + ln (i)) annealed, i = i +1;
5) If the requirement or the number of iterations is met after annealing, the algorithm ends, if f (S) i )<f out Let S stand out p ＝S i+1 Returning to the step 4;
training and predicting a neural network, wherein the training is to determine a network structure by setting fixed input and output, the neural network continuously adjusts the connection weight between each neuron in the training process so as to reduce the error between the training output and the designated output, and the prediction is the process of processing input data by the trained network to obtain output;
step four: and finally, comparing the output result with the form information input by the testee in the experimental process to finish the identification and comparison of the emotional words.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710998545.XA CN107918487A (en) | 2017-10-20 | 2017-10-20 | A kind of method that Chinese emotion word is identified based on skin electrical signal |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710998545.XA CN107918487A (en) | 2017-10-20 | 2017-10-20 | A kind of method that Chinese emotion word is identified based on skin electrical signal |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107918487A true CN107918487A (en) | 2018-04-17 |
Family
ID=61895714
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710998545.XA Pending CN107918487A (en) | 2017-10-20 | 2017-10-20 | A kind of method that Chinese emotion word is identified based on skin electrical signal |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107918487A (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108523906A (en) * | 2018-04-27 | 2018-09-14 | 合肥工业大学 | Personality analysis method and system, storage medium based on pulse characteristics |
| CN108630299A (en) * | 2018-04-27 | 2018-10-09 | 合肥工业大学 | Personality analysis method and system, storage medium based on skin resistance feature |
| CN109344331A (en) * | 2018-10-26 | 2019-02-15 | 南京邮电大学 | A kind of user feeling analysis method based on online community network |
| CN109685149A (en) * | 2018-12-28 | 2019-04-26 | 杭州翼兔网络科技有限公司 | Mood finely incorporates model construction and the automatic method for carrying out mood and finely obtaining into |
| CN109933913A (en) * | 2019-03-15 | 2019-06-25 | 湖南理工学院 | One kind being based on BP neural network high temperature protection clothes optimum design method |
| CN114780693A (en) * | 2022-03-25 | 2022-07-22 | 昆明理工大学 | Deep migration question classification method based on antagonistic learning |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102809718A (en) * | 2012-07-25 | 2012-12-05 | 华南理工大学 | Ultra-high-frequency partial discharge signal identification method for gas insulated switchgear (GIS) |
| CN103892792A (en) * | 2012-12-24 | 2014-07-02 | 中国科学院深圳先进技术研究院 | Emotion recognition model generation device and method |
| CN104887198A (en) * | 2014-03-06 | 2015-09-09 | 中国科学院沈阳自动化研究所 | Pain quantitative analysis system and method based on human body physiological signal multi-parameter fusion |
| CN104965822A (en) * | 2015-07-29 | 2015-10-07 | 中南大学 | Emotion analysis method for Chinese texts based on computer information processing technology |
| CN105930864A (en) * | 2016-04-15 | 2016-09-07 | 杭州电子科技大学 | EEG (electroencephalogram) signal feature classification method based on ABC-SVM |
-
2017
- 2017-10-20 CN CN201710998545.XA patent/CN107918487A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102809718A (en) * | 2012-07-25 | 2012-12-05 | 华南理工大学 | Ultra-high-frequency partial discharge signal identification method for gas insulated switchgear (GIS) |
| CN103892792A (en) * | 2012-12-24 | 2014-07-02 | 中国科学院深圳先进技术研究院 | Emotion recognition model generation device and method |
| CN104887198A (en) * | 2014-03-06 | 2015-09-09 | 中国科学院沈阳自动化研究所 | Pain quantitative analysis system and method based on human body physiological signal multi-parameter fusion |
| CN104965822A (en) * | 2015-07-29 | 2015-10-07 | 中南大学 | Emotion analysis method for Chinese texts based on computer information processing technology |
| CN105930864A (en) * | 2016-04-15 | 2016-09-07 | 杭州电子科技大学 | EEG (electroencephalogram) signal feature classification method based on ABC-SVM |
Non-Patent Citations (2)
| Title |
|---|
| 周钰婷等: "模拟退火免疫粒子群算法在皮肤电信号情感识别中的应用", 《计算机应用》 * |
| 魏先民: "几种改进的人工神经网络算法比较", 《福建电脑》 * |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108523906A (en) * | 2018-04-27 | 2018-09-14 | 合肥工业大学 | Personality analysis method and system, storage medium based on pulse characteristics |
| CN108630299A (en) * | 2018-04-27 | 2018-10-09 | 合肥工业大学 | Personality analysis method and system, storage medium based on skin resistance feature |
| CN109344331A (en) * | 2018-10-26 | 2019-02-15 | 南京邮电大学 | A kind of user feeling analysis method based on online community network |
| CN109685149A (en) * | 2018-12-28 | 2019-04-26 | 杭州翼兔网络科技有限公司 | Mood finely incorporates model construction and the automatic method for carrying out mood and finely obtaining into |
| CN109685149B (en) * | 2018-12-28 | 2021-04-27 | 江苏智慧工场技术研究院有限公司 | Method for constructing emotion fine scoring model and automatically acquiring emotion fine |
| CN109933913A (en) * | 2019-03-15 | 2019-06-25 | 湖南理工学院 | One kind being based on BP neural network high temperature protection clothes optimum design method |
| CN114780693A (en) * | 2022-03-25 | 2022-07-22 | 昆明理工大学 | Deep migration question classification method based on antagonistic learning |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Vásquez-Correa et al. | Multimodal assessment of Parkinson's disease: a deep learning approach | |
| CN107918487A (en) | A kind of method that Chinese emotion word is identified based on skin electrical signal | |
| Rubin et al. | Classifying heart sound recordings using deep convolutional neural networks and mel-frequency cepstral coefficients | |
| CN112656427B (en) | Electroencephalogram emotion recognition method based on dimension model | |
| Zhang et al. | Towards heart sound classification without segmentation using convolutional neural network | |
| CN106073706B (en) | A kind of customized information and audio data analysis method and system towards Mini-mental Status Examination | |
| CN111461176A (en) | Multi-mode fusion method, device, medium and equipment based on normalized mutual information | |
| CN108959895B (en) | Electroencephalogram EEG (electroencephalogram) identity recognition method based on convolutional neural network | |
| CN108960182A (en) | A kind of P300 event related potential classifying identification method based on deep learning | |
| CN113729707A (en) | FECNN-LSTM-based emotion recognition method based on multi-mode fusion of eye movement and PPG | |
| CN105868532B (en) | A kind of method and system of intelligent evaluation heart aging degree | |
| CN113158964B (en) | Sleep stage method based on residual error learning and multi-granularity feature fusion | |
| CN107808663B (en) | Parkinson disease voice data classification system based on DBN and RF algorithm | |
| CN111259895B (en) | Emotion classification method and system based on facial blood flow distribution | |
| CN113705398B (en) | Music electroencephalogram space-time characteristic classification method based on convolution-long and short term memory network | |
| CN113012720A (en) | Depression detection method by multi-voice characteristic fusion under spectral subtraction noise reduction | |
| WO2024098649A1 (en) | Street greening quality testing method based on physiological arousal recognition | |
| CN111000556A (en) | Emotion recognition method based on deep fuzzy forest | |
| CN114998651A (en) | Skin lesion image classification and identification method, system and medium based on transfer learning | |
| CN113674767A (en) | Depression state identification method based on multi-modal fusion | |
| Majda-Zdancewicz et al. | Deep learning vs feature engineering in the assessment of voice signals for diagnosis in Parkinson’s disease | |
| Lu et al. | Speech depression recognition based on attentional residual network | |
| CN114299996A (en) | AdaBoost algorithm-based speech analysis method and system for key characteristic parameters of symptoms of frozen gait of Parkinson's disease | |
| Nouhaila et al. | An intelligent approach based on the combination of the discrete wavelet transform, delta delta MFCC for Parkinson's disease diagnosis | |
| CN114159079A (en) | Multi-type muscle fatigue detection method based on feature extraction and GRU deep learning model |
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 |
Application publication date: 20180417 |
|
| RJ01 | Rejection of invention patent application after publication |