Good Classification Performance Research Articles

Digitization of Civics in College Physical Education Courses Based on the Correlation Matrix

Abstract In this paper, after the extraction of teaching resource features by mutual information method, the classification of Civics teaching resources is realized by constructing a decision tree classifier. According to the current situation of Civics education in university sports courses, the set segmentation correlation function is determined, the contribution of each factor is combined using the correlation matrix, the correlation coefficient weights are calculated, and the design of the Civics multimodal corpus is accomplished based on big data technology. Nine universities in Guangzhou City were selected as the objects of this study, and the research data were obtained through questionnaire surveys of physical education teachers in these nine schools. SPSS26.0 and Origin 2019 software were used to study the Civics and Politics of Physical Education Courses in Colleges and Universities. The results show that the convergence speed of the objective function of the algorithm in this paper and the classification accuracy has been improved substantially, compared with other algorithms, by about 0.3 to 0.7, i.e., it shows that the algorithm in this paper has a good performance of classification of the course Sijian. The scores of national sentiment and sports core literacy are 0.76 and 0.76, respectively, with a rank of medium (less than 0.79), and the score of physical and mental health is 0.82, with a rank of high (greater than or equal to 0.8). That is, the public physical education (volleyball) class has a relatively satisfactory effect on the cultivation of students’ physical and mental health, while there is still room for progress in the cultivation effect of family and national sentiment and sports core literacy. This study supplements, to a certain extent, the lack of an evaluation index system for the effect of sports courses’ ideology and nurturing and promotes the better development of public sports courses’ ideology.

A priori prediction of breast cancer response to neoadjuvant chemotherapy using quantitative ultrasound, texture derivative and molecular subtype

The purpose of this study was to investigate the performances of the tumor response prediction prior to neoadjuvant chemotherapy based on quantitative ultrasound, tumour core-margin, texture derivative analyses, and molecular parameters in a large cohort of patients (n = 208) with locally advanced and earlier-stage breast cancer and combined them to best determine tumour responses with machine learning approach. Two multi-features response prediction algorithms using a k-nearest neighbour and support vector machine were developed with leave-one-out and hold-out cross-validation methods to evaluate the performance of the response prediction models. In a leave-one-out approach, the quantitative ultrasound-texture analysis based model attained good classification performance with 80% of accuracy and AUC of 0.83. Including molecular subtype in the model improved the performance to 83% of accuracy and 0.87 of AUC. Due to limited number of samples in the training process, a model developed with a hold-out approach exhibited a slightly higher bias error in classification performance. The most relevant features selected in predicting the response groups are core-to-margin, texture-derivative, and molecular subtype. These results imply that that baseline tumour-margin, texture derivative analysis methods combined with molecular subtype can potentially be used for the prediction of ultimate treatment response in patients prior to neoadjuvant chemotherapy.

Smoking Classification Using Novel Plasma Cytokines by implementing Machine Learning and Statistical Methods.

Smoking is a major cause of premature and preventable death. Tobacco exposure has a detrimental effect on many organs and contributes to multiple diseases including chronic obstructive pulmonary disease (COPD), cardiovascular disease, cancer, and diabetes. Cytokines are inflammatory biomarkers that are mechanistically associated with smoking. Machine Learning algorithms allow for the quantitative assessment of the contributions of individual cytokines to tobacco-related diseases. The mapping of cytokines to disease can facilitate and direct treatment modalities. By the application of k Nearest Neighbor (k-NN) and Random Forest machine learning algorithms on 63 plasma cytokines we have demonstrated the classification of smoking. To ensure optimal results, performance improvement techniques such as k-fold cross validation and hyper parameter tuning are employed. Separability efficiency achieved by the models is evaluated using the Area Under the Receiver Operating Characteristic (AUROC) metric. The most significant cytokines that enabled the classification are identified and presented. The statistically significant difference for AUROC score of k-NN and Random Forest has been ascertained using the 2-sample independent t test. A reasonably good classification performance was achieved by k-NN algorithm with an AUROC metric of .87, and a 95% CI of (.823,.917). Random forest exceeded k-NN algorithm's performance, with a perfect AUROC score of 1 and a 95% CI of (1,1). From among the ten most prominent cytokines that contributed to the classification, the ones common to both algorithms are: LIF, IL22, G-CSF/CSF-3, TRAIL. AUROC scores for k-NN and Random Forest are significantly different (p-value = 5.105e-16). The discovery and transference of biomarkers such as cytokines from the platform of molecular investigation to clinical practice, can facilitate precision medicine-based therapeutic interventions.

Colorimetry/fluorescence dual-mode detection of Salmonella typhimurium based on self-assembly of MCOF with Au NPs nanozyme coupled AIEgen

Sensitive, accurate, simple and quick monitoring of Salmonella typhimurium (S. typhimurium) in food is significant for preventing food poisoning, but still remains a challenge. Herein, a colorimetry/fluorescence dual-mode sensing strategy was fabricated to detect S. typhimurium by integrating the self-assembly of magnetic covalent organic framework (MCOF) with gold nanoparticles (Au NPs) as the peroxidase-mimicking nanozyme and aggregation-induced emission luminogen (AIEgen). S. typhimurium could competitive bind to aptamer conjugated Au NPs (Au NPs@apt), inhibit the self-assembly of MCOF with Au NPs, and shield the catalytic activity of AuNPs. After adding H2O2 and TPE-4A, the dark green solution changed to light with increasing S. typhimurium concentration, on the contrary, the fluorescent signals were generated. As a result, in colorimetry/fluorescence modes, S. typhimurium could be detected in the linear ranges of 103–108 CFU mL−1 and 101–107 CFU mL−1, with LODs of 1000 and 10 CFU mL−1, respectively. Importantly, different colors consistent with various S. typhimurium concentrations can be accurately classified by smartphone app and linear discriminant analysis (LDA). The smartphone-assisted data interpretation can generate complementary colorimetry and fluorescence signals without any sophisticated equipment and achieve on-site detection. Moreover, the proposed strategy could be explored for S. typhimurium monitoring in milk with satisfactory recoveries (97.6–100.4 %) in colorimetry and fluorescence mode and good classification and prediction performance in smartphone/LDA system, suggesting the feasibility and potential applications of the sensing platform.

Deep Learning of Diabetic Retinopathy Classification in Fundus Images

Diabetic retinopathy is an eye disease in diabetic patients due to damage to the small blood vessels in the retina due to high and low blood sugar levels. Accurate detection and classification of Diabetic Retinopathy is an important task in computer-aided diagnosis, especially when planning for diabetic retinopathy surgery. Therefore, this study aims to design an automated model based on deep learning, which helps ophthalmologists detect and classify diabetic retinopathy severity through fundus images. In this work, a deep convolutional neural network (CNN) with transfer learning and fine tunes has been proposed by using pre-trained networks known as Residual Network-50 (ResNet-50). The overall framework of the proposed classification model is divided into three major phases, including pre-processing, training the Resnet-50 network, and classification with evaluation. In the first phase, pre-processing techniques are applied to the APTOS2019 fundus images dataset to find the best features and highlight some fine details of these images. The resnet-50 network was trained in the second phase using the training set and saved the best model obtained that gives high accuracy during the training process. Finally, this saved model has been implemented on the testing dataset for classification DR grades. The proposed model shows good and best classification performance, which was obtained with an accuracy of 98.3%, a precision of 98.4%, an F1-Score of 98.5 % and the recall of 98.4%.

Improving Brain Tumor Classification Efficacy through the Application of Feature Selection and Ensemble Classifiers

Accurate brain tumor detection is crucial due to its high mortality rate. However, existing automated methods suffer from limited accuracy and high false-positive rates. In this study, we aimed to improve brain tumor classification by comparing 17 different classifiers organized into six groups: Decision Tree (DT) Model, Support Vector Machine (SVM), Naive Bayes Classifier, Logistic Regression, Generalized Linear Model (GLM) Classifier, and Neural Network. We utilized a dataset of 3,762 Magnetic Resonance Imaging (MRI) scans of brain tumors from Kaggle, with each image having dimensions of 240 × 240 pixels and labeled as tumor or non-tumor. Our approach involved three main steps: extracting visual information using 17 predictor classes, optimizing feature extraction through weight optimization, and comparing different sets of classifier models. We evaluated the models’ performance using the confusion matrix and Receiver Operating Characteristics (ROC) curves. Our results showed that optimizing feature selection and utilizing ensemble classifiers improved the accuracy of brain tumor classification. The DT Model with ensemble classifiers emerged as the best-performing classifier, achieving an accuracy of 98.11% and an AUC of 0.99. Notably, Random Tree (RT) exhibited the highest accuracy within the ensemble classifier set, with a significant increase compared to other models. Our proposed method outperformed the standard approach, demonstrating its potential for enhancing brain tumor detection accuracy. This study contributes to the field by providing a more accurate method for detecting brain tumors, potentially enabling earlier detection and improved patient outcomes. Future research should focus on further improving brain tumor diagnosis and treatment through the application of machine learning techniques.

A Novel Multiscale Dilated Convolution Neural Network With Gating Mechanism for Decoding Driving Intentions Based on EEG

Deep learning methods based on convolution neural networks (CNNs) have achieved good classification performance in decoding electroencephalography (EEG). In this paper, a novel framework combining gating mechanism and dilated convolution neural network (GDCNN) is proposed for decoding EEG signals evoked by four different driving intentions. GDCNN provides different receptive fields and controls the information flow between convolution layers, which help to detect different sizes of information in EEG signals. The proposed method reaches accuracies of 93.17% and 73.33% for the subject-dependent and subject-independent experiments, outperforming several benchmark methods. The data augmentation that randomly concatenates multi-trial EEG sequences is adopted to promote generalization of decoding model. This strategy effectively prevents overfitting and improves the decoding accuracies of EEGNet, DeepConvNet, and GDCNN by 4.3%, 4.75%, and 3.92%, respectively. These results indicate GDCNN is beneficial for decoding EEG and it has application potential in the brain-computer interface (BCI) systems based on video stimuli.

MLDA: Multi-Loss Domain Adaptor for Cross-Session and Cross-Emotion EEG-Based Individual Identification.

Traditional individual identification methods, such as face and fingerprint recognition, carry the risk of personal information leakage. The uniqueness and privacy of electroencephalograms (EEG) and the popularization of EEG acquisition devices have intensified research on EEG-based individual identification in recent years. However, most existing work uses EEG signals from a single session or emotion, ignoring large differences between domains. As EEG signals do not satisfy the traditional deep learning assumption that training and test sets are independently and identically distributed, it is difficult for trained models to maintain good classification performance for new sessions or new emotions. In this article, an individual identification method, called Multi-Loss Domain Adaptor (MLDA), is proposed to deal with the differences between marginal and conditional distributions elicited by different domains. The proposed method consists of four parts: a) Feature extractor, which uses deep neural networks to extract deep features from EEG data; b) Label predictor, which uses full-layer networks to predict subject labels; c) Marginal distribution adaptation, which uses maximum mean discrepancy (MMD) to reduce marginal distribution differences; d) Associative domain adaptation, which adapts to conditional distribution differences. Using the MLDA method, the cross-session and cross-emotion EEG-based individual identification problem is addressed by reducing the influence of time and emotion. Experimental results confirmed that the method outperforms other state-of-the-art approaches.

Multilevel Laser-Induced Pain Measurement with Wasserstein Generative Adversarial Network- Gradient Penalty Model.

Pain is an experience of unpleasant sensations and emotions associated with actual or potential tissue damage. In the global context, billions of people are affected by pain disorders. There are particular challenges in the measurement and assessment of pain, and the commonly used pain measuring tools include traditional subjective scoring methods and biomarker-based measures. The main tools for biomarker-based analysis are electroencephalography (EEG), electrocardiography and functional magnetic resonance. The EEG-based quantitative pain measurements are of immense value in clinical pain management and can provide objective assessments of pain intensity. The assessment of pain is now primarily limited to the identification of the presence or absence of pain, with less research on multilevel pain. High power laser stimulation pain experimental paradigm and five pain level classification methods based on EEG data augmentation are presented. First, the EEG features are extracted using modified S-transform, and the time-frequency information of the features is retained. Based on the pain recognition effect, the 20-40[Formula: see text]Hz frequency band features are optimized. Afterwards the Wasserstein generative adversarial network with gradient penalty is used for feature data augmentation. It can be inferred from the good classification performance of features in the parietal region of the brain that the sensory function of the parietal lobe region is effectively activated during the occurrence of pain. By comparing the latest data augmentation methods and classification algorithms, the proposed method has significant advantages for the five-level pain dataset. This research provides new ways of thinking and research methods related to pain recognition, which is essential for the study of neural mechanisms and regulatory mechanisms of pain.

Aflatoxins detection in almonds via fluorescence imaging and deep neural network approach

The detection of aflatoxin B in raw food materials represents a topic of great interest worldwide because of the huge health and economic impact of aflatoxin contamination. In this paper, we present an original approach to aflatoxin detection, using a portable instrument to acquire fluorescence images, among other spectral responses. The acquired images are processed by combining a color space conversion from the RGB scale to Y′CbCr, and a neural network approach to encode a vector of features. After a feature reduction using a Receiving Operating Curve method, two-class and three-class classification tasks of contaminated vs non-contaminated samples are accomplished. This procedure has been applied to artificially contaminated grained almond samples in the range of 0–320.2 ng/g, achieving an overall classification accuracy between 84.7% and 93.0%, depending on the number of classes. Thus, in this setting, we show that good classification performance can be achieved using only an image acquisition and analysis approach. The proposed procedure can represent a cheap, rapid, non-destructive yet sensitive method for the assessment of aflatoxin B contamination in food matrices, and its monitoring and tracing throughout the food chain.

Automatic Quantification of Normal Brain Gyrification Patterns and Changes in Fetuses with Polymicrogyria and Lissencephaly Based on MRI.

The current imaging assessment of fetal brain gyrification is performed qualitatively and subjectively using sonography and MR imaging. A few previous studies have suggested methods for quantification of fetal gyrification based on 3D reconstructed MR imaging, which requires unique data and is time-consuming. In this study, we aimed to develop an automatic pipeline for gyrification assessment based on routinely acquired fetal 2D MR imaging data, to quantify normal changes with gestation, and to measure differences in fetuses with lissencephaly and polymicrogyria compared with controls. We included coronal T2-weighted MR imaging data of 162 fetuses retrospectively collected from 2 clinical sites: 134 controls, 12 with lissencephaly, 13 with polymicrogyria, and 3 with suspected lissencephaly based on sonography, yet with normal MR imaging diagnoses. Following brain segmentation, 5 gyrification parameters were calculated separately for each hemisphere on the basis of the area and ratio between the contours of the cerebrum and its convex hull. Seven machine learning classifiers were evaluated to differentiate control fetuses and fetuses with lissencephaly or polymicrogyria. In control fetuses, all parameters changed significantly with gestational age (P < .05). Compared with controls, fetuses with lissencephaly showed significant reductions in all gyrification parameters (P ≤ .02). Similarly, significant reductions were detected for fetuses with polymicrogyria in several parameters (P ≤ .001). The 3 suspected fetuses showed normal gyrification values, supporting the MR imaging diagnosis. An XGBoost-linear algorithm achieved the best results for classification between fetuses with lissencephaly and control fetuses (n = 32), with an area under the curve of 0.90 and a recall of 0.83. Similarly, a random forest classifier showed the best performance for classification of fetuses with polymicrogyria and control fetuses (n = 33), with an area under the curve of 0.84 and a recall of 0.62. This study presents a pipeline for automatic quantification of fetal brain gyrification and provides normal developmental curves from a large cohort. Our method significantly differentiated fetuses with lissencephaly and polymicrogyria, demonstrating lower gyrification values. The method can aid radiologic assessment, highlight fetuses at risk, and may improve early identification of fetuses with cortical malformations.

Analysis of pipe sticking due to wellbore uncleanliness using machine learning

Pipe sticking is a major problem that can result in significant downtime, lost productivity, and increased costs. The aim of this study is to develop an accurate and effective predictive model for pipe sticking due to wellbore uncleanliness using a range of classification algorithms. In this research work, the drilling data from two different reservoirs in India was pre-processed and eviscerated to ensure that it was suitable to process using classification algorithms. The data collected from two different reservoirs in India were analyzed using different machine learning algorithms to tackle the persistent challenges of pipe sticking during oil drilling operations. These algorithms were compared and evaluated based on their performance. The research finding indicates that the ensemble classifier algorithm performs better than a single classifier algorithm. It shows high generalization ability with an average accuracy of around 90 %. In addition to this, the ensemble classifier algorithm possesses good classification performance, and provides immunity from noisy data, offering strong support for real-time detection to prevent pipe sticking thereby reducing costly downtime.

Multistream BertGCN for Sentiment Classification Based on Cross-Document Learning

Very recently, the BERT graph convolutional network (BertGCN) model has attracted much attention from researchers due to its good text classification performance. However, just using original documents in the corpus to construct the topology of graphs for GCN-based models may lose some effective information. In this paper, we focus on sentiment classification, an important branch of text classification, and propose the multistream BERT graph convolutional network (MS-BertGCN) for sentiment classification based on cross-document learning. In the proposed method, we first combine the documents in the training set based on within-class similarity. Then, each heterogeneous graph is constructed using a group of combinations of documents for the single-stream BertGCN model. Finally, we construct multistream-BertGCN (MS-BertGCN) based on multiple heterogeneous graphs constructed from different groups of combined documents. The experimental results show that our MS-BertGCN model outperforms state-of-the-art methods on sentiment classification tasks.

Training a Hyperdimensional Computing Classifier Using a Threshold on Its Confidence.

Hyperdimensional computing (HDC) has become popular for light-weight and energy-efficient machine learning, suitable for wearable Internet-of-Things devices and near-sensor or on-device processing. HDC is computationally less complex than traditional deep learning algorithms and achieves moderate to good classification performance. This letter proposes to extend the training procedure in HDC by taking into account not only wrongly classified samples but also samples that are correctly classified by the HDC model but with low confidence. We introduce a confidence threshold that can be tuned for each data set to achieve the best classification accuracy. The proposed training procedure is tested on UCIHAR, CTG, ISOLET, and HAND data sets for which the performance consistently improves compared to the baseline across a range of confidence threshold values. The extended training procedure also results in a shift toward higher confidence values of the correctly classified samples, making the classifier not only more accurate but also more confident about its predictions.

Skin cancer classification using improved transfer learning model‐based random forest classifier and golden search optimization

AbstractThe two complex skin cancer type is melanoma and non‐melanoma. However, the classification of skin cancer is not performed accurately due to the requirement of a huge time and the presence of noisy images. The severity level is not detected and this leads to generate serious issues. To overcome these entire issues GSO‐optimized kernel random forest (RF)‐based transfer learning (TL) is proposed to perform the classification of 12 different skin cancer types by using ImageNet and CNN architectures. The kernel decision tree is taken into account by the RF classifier throughout the classification process, and the RF parameters are modified using the GSO algorithm. The proposed algorithm not only effectively enhances the detection results based on previous adaptive thresholding segmentation techniques but also extends the skin cancer detection process. Compared to existing MMCL, SVM, Mask‐RCNN, and CNN methods the proposed technique attained better performance in classification of skin cancer.

3D-2D Distance Maps Conversion Enhances Classification of Craniosynostosis.

Diagnosis of craniosynostosis using photogrammetric 3D surface scans is a promising radiation-free alternative to traditional computed tomography. We propose a 3D surface scan to 2D distance map conversion enabling the usage of the first convolutional neural networks (CNNs)-based classification of craniosynostosis. Benefits of using 2D images include preserving patient anonymity, enabling data augmentation during training, and a strong under-sampling of the 3D surface with good classification performance. The proposed distance maps sample 2D images from 3D surface scans using a coordinate transformation, ray casting, and distance extraction. We introduce a CNNbased classification pipeline and compare our classifier to alternative approaches on a dataset of 496 patients. We investigate into low-resolution sampling, data augmentation, and attribution mapping. Resnet18 outperformed alternative classifiers on our dataset with an F1-score of 0.964 and an accuracy of 98.4 %. Data augmentation on 2D distance maps increased performance for all classifiers. Under-sampling allowed 256-fold computation reduction during ray casting while retaining an F1-score of 0.92. Attribution maps showed high amplitudes on the frontal head. We demonstrated a versatile mapping approach to extract a 2D distance map from the 3D head geometry increasing classification performance, enabling data augmentation during training on 2D distance maps, and the usage of CNNs. We found that low-resolution images were sufficient for a good classification performance. Photogrammetric surface scans are a suitable craniosynostosis diagnosis tool for clinical practice. Domain transfer to computed tomography seems likely and can further contribute to reducing ionizing radiation exposure for infants.

A Novel Semantic Segmentation Approach Using Improved SegNet and DSC in Remote Sensing Images

An improved SegNet semantic segmentation model is proposed to address the issue of traditional classification algorithms and shallow learning algorithms not being suitable for extracting information from high-resolution remote sensing images. During the research process, space remote sensing images obtained from the GF-1 satellite were used as the data source. In order to improve the operational efficiency of the encoding network, the pooling layer in the encoding network is removed and the ordinary convolutional layer is replaced with a depth-wise separable convolution. By decoding the last layer of the network to obtain the reshaped output results, and then calculating the probability of each classification using a Softmax classifier, the classification of pixels can be achieved. The output result of the classifier is the final result of the remote sensing image semantic segmentation model. The results showed that the proposed algorithm had the highest Kappa coefficient of 0.9531, indicating good classification performance.

The Jena Audiovisual Stimuli of Morphed Emotional Pseudospeech (JAVMEPS): A database for emotional auditory-only, visual-only, and congruent and incongruent audiovisual voice and dynamic face stimuli with varying voice intensities.

We describe JAVMEPS, an audiovisual (AV) database for emotional voice and dynamic face stimuli, with voices varying in emotional intensity. JAVMEPS includes 2256 stimulus files comprising (A) recordings of 12speakers, speaking four bisyllabic pseudowords with six naturalistic induced basic emotions plus neutral, in auditory-only, visual-only, and congruent AV conditions. It furthermore comprises (B) caricatures (140%), original voices (100%), and anti-caricatures (60%) for happy, fearful, angry, sad, disgusted, and surprised voices for eight speakers and two pseudowords. Crucially, JAVMEPS contains (C) precisely time-synchronized congruent and incongruent AV (and corresponding auditory-only) stimuli with two emotions (anger, surprise), (C1) with original intensity (ten speakers, four pseudowords), (C2) and with graded AV congruence (implemented via five voice morph levels, from caricatures to anti-caricatures; eight speakers, two pseudowords). We collected classification data for Stimulus Set A from 22 normal-hearing listeners and four cochlear implant users, for two pseudowords, in auditory-only, visual-only, and AV conditions. Normal-hearing individuals showed good classification performance (McorrAV = .59 to .92), with classification rates in the auditory-only condition ≥ .38 correct (surprise: .67, anger: .51). Despite compromised vocal emotion perception, CI users performed above chance levels of .14 for auditory-only stimuli, with best rates for surprise (.31) and anger (.30). We anticipate JAVMEPS to become a useful open resource for researchers into auditory emotion perception, especially when adaptive testing or calibration of task difficulty is desirable. With its time-synchronized congruent and incongruent stimuli, JAVMEPS can also contribute to filling a gap in research regarding dynamic audiovisual integration of emotion perception via behavioral or neurophysiological recordings.

A Framework for Synthetic Agetech Attack Data Generation

To address the lack of datasets for agetech, this paper presents an approach for generating synthetic datasets that include traces of benign and attack datasets for agetech. The generated datasets could be used to develop and evaluate intrusion detection systems for smart homes for seniors aging in place. After reviewing several resources, it was established that there are no agetech attack data for sensor readings. Therefore, in this research, several methods for generating attack data were explored using attack data patterns from an existing IoT dataset called TON_IoT weather data. The TON_IoT dataset could be used in different scenarios, but in this study, the focus is to apply it to agetech. The attack patterns were replicated in a normal agetech dataset from a temperature sensor collected from the Information Security and Object Technology (ISOT) research lab. The generated data are different from normal data, as abnormal segments are shown that could be considered as attacks. The generated agetech attack datasets were also trained using machine learning models, and, based on different metrics, achieved good classification performance in predicting whether a sample is benign or malicious.

Research on face emotion recognition algorithm based on deep learning neural network

Abstract This paper explores face emotion recognition based on deep learning neural networks in order to make the computer recognize facial emotion more accurately. Firstly, we introduce the convolutional neural network, which consists of a convolutional layer, pooling layer and fully connected layer in the deep learning neural network, and propose the variational self-encoder face emotion recognition algorithm based on the convolutional neural network in combination with the facial emotion recognition process. The correct recognition rates achieved by EVA using 32×32 pixels, LBP features and 50-dimensional Gabor wavelet features are 95.13%, 96.74% and 98.84%, respectively. In terms of emotion recognition ability, the recognition accuracy percentages were around 70% for large emotion fluctuations and around 30% for small emotion fluctuations. Since the neutral facial emotion is not well distinguished, the recognition degree of neutral facial emotion is only 80%, and the recognition degree using the self-encoder-based face emotion recognition algorithm in several facial features extraction is above 80%, and the recognition degree with neutral emotion removed is up to about 90%. Therefore, the algorithm in this paper has good classification performance and can recognize and analyze facial emotion features accurately.