Terms Of Classification Accuracy Research Articles

MTLBORKS-CNN: An Innovative Approach for Automated Convolutional Neural Network Design for Image Classification

Convolutional neural networks (CNNs) have excelled in artificial intelligence, particularly in image-related tasks such as classification and object recognition. However, manually designing CNN architectures demands significant domain expertise and involves time-consuming trial-and-error processes, along with substantial computational resources. To overcome this challenge, an automated network design method known as Modified Teaching-Learning-Based Optimization with Refined Knowledge Sharing (MTLBORKS-CNN) is introduced. It autonomously searches for optimal CNN architectures, achieving high classification performance on specific datasets without human intervention. MTLBORKS-CNN incorporates four key features. It employs an effective encoding scheme for various network hyperparameters, facilitating the search for innovative and valid network architectures. During the modified teacher phase, it leverages a social learning concept to calculate unique exemplars that effectively guide learners while preserving diversity. In the modified learner phase, self-learning and adaptive peer learning are incorporated to enhance knowledge acquisition of learners during CNN architecture optimization. Finally, MTLBORKS-CNN employs a dual-criterion selection scheme, considering both fitness and diversity, to determine the survival of learners in subsequent generations. MTLBORKS-CNN is rigorously evaluated across nine image datasets and compared with state-of-the-art methods. The results consistently demonstrate MTLBORKS-CNN’s superiority in terms of classification accuracy and network complexity, suggesting its potential for infrastructural development of smart devices.

An Efficient Binary Sand Cat Swarm Optimization for Feature Selection in High-Dimensional Biomedical Data.

Recent breakthroughs are making a significant contribution to big data in biomedicine which are anticipated to assist in disease diagnosis and patient care management. To obtain relevant information from this data, effective administration and analysis are required. One of the major challenges associated with biomedical data analysis is the so-called "curse of dimensionality". For this issue, a new version of Binary Sand Cat Swarm Optimization (called PILC-BSCSO), incorporating a pinhole-imaging-based learning strategy and crossover operator, is presented for selecting the most informative features. First, the crossover operator is used to strengthen the search capability of BSCSO. Second, the pinhole-imaging learning strategy is utilized to effectively increase exploration capacity while avoiding premature convergence. The Support Vector Machine (SVM) classifier with a linear kernel is used to assess classification accuracy. The experimental results show that the PILC-BSCSO algorithm beats 11 cutting-edge techniques in terms of classification accuracy and the number of selected features using three public medical datasets. Moreover, PILC-BSCSO achieves a classification accuracy of 100% for colon cancer, which is difficult to classify accurately, based on just 10 genes. A real Liver Hepatocellular Carcinoma (TCGA-HCC) data set was also used to further evaluate the effectiveness of the PILC-BSCSO approach. PILC-BSCSO identifies a subset of five marker genes, including prognostic biomarkers HMMR, CHST4, and COL15A1, that have excellent predictive potential for liver cancer using TCGA data.

Potato Plant Leaf Disease Detection Using Deep Learning Method

In agriculture, plant disease detection is an important concern to achieve high crop production and yield sustainably. Automated detection and analysis could be beneficial for early action to prevent spreading, cure the plant in earlier stages, reduce the damage, and protect crop or forest health. This study proposes a new deep-learning model that correctly classifies plant leaf diseases for the agriculture and food sectors. It focuses on the detection of plant diseases for potato leaves from images by designing a new convolutional neural network architecture. The experimental results conducted on a real-world dataset showed that a significant improvement (8.6%) was achieved on average by the proposed model (98.28%) compared to the state-of-the-art models (89.67%) in terms of classification accuracy.

A comparative study of ensemble methods in the field of education: Bagging and Boosting algorithms

This study aims to conduct a comparative study of Bagging and Boosting algorithms among ensemble methods and to compare the classification performance of TreeNet and Random Forest methods using these algorithms on the data extracted from ABİDE application in education. The main factor in choosing them for analyses is that they are Ensemble methods combining decision trees via Bagging and Boosting algorithms and creating a single outcome by combining the outputs obtained from each of them. The data set consists of mathematics scores of ABİDE (Academic Skills Monitoring and Evaluation) 2016 implementation and various demographic variables regarding students. The study group involves 5000 students randomly recruited. On the deletion of loss data and assignment procedures, this number decreased to 4568. The analyses showed that the TreeNet method performed more successfully in terms of classification accuracy, sensitivity, F1-score and AUC value based on sample size, and the Random Forest method on specificity and accuracy. It can be alleged that the TreeNet method is more successful in all numerical estimation error rates for each sample size by producing lower values compared to the Random Forest method. When comparing both analysis methods based on ABİDE data, considering all the conditions, including sample size, cross validity and performance criteria following the analyses, TreeNet can be said to exhibit higher classification performance than Random Forest. Unlike a single classifier or predictive method, the classification or prediction of multiple methods by using Boosting and Bagging algorithms is considered important for the results obtained in education.

Copula entropy-based golden jackal optimization algorithm for high-dimensional feature selection problems

Feature selection (FS) is a crucial process that aims to remove unnecessary features from datasets. It plays a role in data mining and machine learning (ML) by reducing the risk associated with high-dimensional datasets. FS is considered a challenging problem that is difficult to solve efficiently due to its combinatorial nature. As the size of the problem increases, the computation time also grows. Recently, researchers have focused on metaheuristic FS algorithms specifically designed for high-dimensional datasets. Therefore, this article proposes a powerful metaheuristic algorithm called Binary Enhanced Golden Jackal Optimization (BEGJO), which is an improved version of the recently published Golden Jackal Optimization (GJO) algorithm. The original GJO algorithm faces challenges when dealing with high-dimensional FS problems, as it tends to get trapped in local optima. To address this issue, various enhancement strategies are employed to improve the efficiency of GJO. The proposed BEGJO algorithm utilizes Copula Entropy (CE) to reduce the dimensionality of high-dimensional FS problems while maintaining high classification accuracy using the K-Nearest Neighbour (K-NN) classifier. Additionally, four enhancement strategies are incorporated to enhance the exploration and exploitation capabilities of the fundamental GJO algorithm. The BEGJO algorithm is transformed into its binary form using the sigmoid transfer function, aligning it with the nature of the FS problem. It is then tested on various high-dimensional benchmark datasets. The effectiveness of BEGJO is evaluated by comparing it with well-known algorithms in terms of classification accuracy, feature dimension, and processing time. BEGJO outperforms other algorithms in terms of classification accuracy and feature dimension and ranks up to fourth in terms of processing time. Furthermore, the advantageous use of CE is demonstrated by comparing the performance of the proposed algorithm with traditional FS algorithms. Statistical evaluations are conducted to further validate the effectiveness and superiority of the proposed algorithm. The results confirm that BEGJO is an effective solution for high-dimensional FS problems.

Research on learning behavior patterns from the perspective of educational data mining: Evaluation, prediction and visualization

The rapid growth of educational data creates the requirement to mine useful information from learning behavior patterns. The development of data mining technology makes educational data mining possible. The paper intends to use a public educational data set to study learning behavior patterns from the perspective of educational data mining, so as to promote the innovation of educational management. Firstly, in order to reduce the dimension of data analysis that facilitates the improvement in efficiency, principal component analysis is carried out to reduce the number of attributes in the data set. The significant attributes in the rotating principal component matrix rather than principal components which are not closely related to learning behavior patterns are extracted as the research variables. Then, a pseudo statistic is proposed to determine the number of clusters and the preprocessed data set is clustered according to the extracted attributes. The clustering results are applied to add class labels to the data, which is convenient for the later data training. Finally, six classification algorithms J48, K-Nearest Neighbor, Bayes Net, Random Forest, Support Vector Machine and Logit Boost are used to train the data with labels and build prediction models. At the same time, the performance and applicable conditions of six classifiers in terms of accuracy, efficiency, error, and so on are discussed and compared. It is found that the performance of the integrated algorithm is better than that of a single classifier. In the integrated algorithm, compared with Random Forest, the running time of Logit Boost is shorter.

Advancing Healthcare: Synergizing Biosensors and Machine Learning for Early Cancer Diagnosis.

Cancer is a fatal disease and a significant cause of millions of deaths. Traditional methods for cancer detection often have limitations in identifying the disease in its early stages, and they can be expensive and time-consuming. Since cancer typically lacks symptoms and is often only detected at advanced stages, it is crucial to use affordable technologies that can provide quick results at the point of care for early diagnosis. Biosensors that target specific biomarkers associated with different types of cancer offer an alternative diagnostic approach at the point of care. Recent advancements in manufacturing and design technologies have enabled the miniaturization and cost reduction of point-of-care devices, making them practical for diagnosing various cancer diseases. Furthermore, machine learning (ML) algorithms have been employed to analyze sensor data and extract valuable information through the use of statistical techniques. In this review paper, we provide details on how various machine learning algorithms contribute to the ongoing development of advanced data processing techniques for biosensors, which are continually emerging. We also provide information on the various technologies used in point-of-care cancer diagnostic biosensors, along with a comparison of the performance of different ML algorithms and sensing modalities in terms of classification accuracy.

Modulation recognition of optical and electromagnetic waves based on transfer learning

Light is a type of electromagnetic wave. Research on the modulation recognition of communication signals, which are also electromagnetic waves, can be applied to the field of modulation and demodulation of optical waves. The limited signal data and varying data distribution have an impact on the automatic modulation recognition of signals. To address this, we proposed a method to enhance the characteristic information of modulation patterns carried by signal data. We used the Wigner–Ville distribution and short-time Fourier transform to convert the original I/Q multi-channel modulated communication signal into time–frequency representation images (TFRIs). Additionally, we adopted a transfer learning approach, the domain adaptation network (DAN), to identify modulation modes of signals under different SNRs. The combination of signal conversion and transfer learning significantly improved the identification of modulation recognition. Our experiments demonstrated that our proposed method outperformed existing state-of-the-art methods in terms of classification accuracy at both high and low signal-to-noise ratios (SNRs). Specifically, our method achieved a 10% increase in Top1-accuracy under high SNR, and overall improved the classification accuracy under low SNR compared to the adversarial transfer learning architecture (ATLA). This is attributed to the use of a more diverse and accurate dataset that provides a better representation of real-world communication signals. We also compared our experimental results with those of other papers, and the results show that our method achieved the best performance among the current state-of-the-art methods, considering the classification accuracy at different SNRs.

Comparing MIMIC and MIMIC-interaction to Alignment Methods for Investigating Measurement Invariance concerning a Continuous Violator

Measurement invariance holds when a latent construct is measured in the same way across different levels of background variables (continuous or categorical) while controlling for the true value of that construct. Using Monte Carlo simulation, this paper compares the multiple indicators, multiple causes (MIMIC) model and MIMIC-interaction to a novel use of alignment optimization (AO) for detecting measurement noninvariance when the violator is a continuous variable. Results showed that MIMIC and MIMIC-interaction in sequential likelihood ratio tests and Wald tests with a Bonferroni correction provided a good balance between identifying invariant and noninvariant (linear violations) items when n ≥ 500 in terms of classification accuracy (CA). AO (CA ≥ .86) was as competitive as MIMIC and MIMIC-interaction to linear invariance violations but was far better under nonlinear quadratic violations when n ≥ 1,000 (i.e., 100 per group for 10 groups).

Classification of Motor Imagery Using Trial Extension in Spatial Domain with Rhythmic Components of EEG

Electrical activities of the human brain can be recorded with electroencephalography (EEG). To characterize motor imagery (MI) tasks for brain–computer interface (BCI) implementation is an easy and cost-effective tool. The MI task is represented by a short-time trial of multichannel EEG. In this paper, the signal of each channel of raw EEG is decomposed into a finite set of narrowband signals using a Fourier-transformation-based bandpass filter. Rhythmic components of EEG are represented by each of the narrowband signals that characterize the brain activities related to MI tasks. The subband signals are arranged to extend the dimension of the EEG trial in the spatial domain. The spatial features are extracted from the set of extended trials using a common spatial pattern (CSP). An optimum number of features are employed to classify the motor imagery tasks using an artificial neural network. An integrated approach with full-band and narrowband signals is implemented to derive discriminative features for MI classification. In addition, the subject-dependent parameter optimization scheme enhances the performance of the proposed method. The performance evaluation of the proposed method is obtained using two publicly available benchmark datasets (Dataset I and Dataset II). The experimental results in terms of classification accuracy (93.88% with Dataset I and 91.55% with Dataset II) show that it performs better than the recently developed algorithms. The enhanced MI classification accuracy is very much applicable in BCI implementation.

Transfer Learning for Automatic Modulation Recognition Using a Few Modulated Signal Samples

This letter proposes a transfer learning model for automatic modulation recognition (AMR) with only a few modulated signal samples. The transfer model is trained with the audio signal UrbanSound8K as the source domain, and then fine-tuned with a few modulated signal samples as the target domain. For improving the classification performance, the signal-to-noise ratio (SNR) is utilized as a feature to facilitate the classification of signals. Simulation results indicate that the transfer model has a significant superiority in terms of classification accuracy.

ABT-SVDD: A method for uncertainty handling in domain adaptation using belief function theory

Domain adaptation involves adapting a model trained on one domain to work effectively on another, which can have different statistical properties, such as distributions, correlations, and relationships between features. These heterogeneities can lead to uncertainty, impacting the model’s performance. Despite many studies that have been done on domain adaptation, most have ignored the adverse impact of uncertain and noisy data on adaptation and classification. To address this issue, the proposed method, Adaptive Belief-based Twin Support Vector Data Description (ABT-SVDD), extends the one-class support vector data description (SVDD) to an adaptive twin classifier and integrates it with a belief-based sample weighting approach. Also, it utilizes a combination of Hermite polynomial and Gaussian kernels to enhance the computational power of the linear objective function of the SVDD classifier while improving the generalization capability. The effectiveness of ABT-SVDD has been compared to the state-of-the-art methods on several tasks taken from two benchmark datasets. The experimental results demonstrate that ABT-SVDD significantly improves classification accuracy on various tasks with varying amounts of labeled data in the target domain. Specifically, in normal situations, ABT-SVDD outperforms competing methods by 6.33% to 9.08%, while in noisy situations, it achieves a more significant improvement of 9.87% compared to the competing methods. Besides, the Wilcoxon statistical test demonstrates the superiority of ABT-SVDD over state-of-the-art ones in terms of classification accuracy and computational time.

Spectral Segmentation Multi-Scale Feature Extraction Residual Networks for Hyperspectral Image Classification

Hyperspectral image (HSI) classification is a vital task in hyperspectral image processing and applications. Convolutional neural networks (CNN) are becoming an effective approach for categorizing hyperspectral remote sensing images as deep learning technology advances. However, traditional CNN usually uses a fixed kernel size, which limits the model’s capacity to acquire new features and affects the classification accuracy. Based on this, we developed a spectral segmentation-based multi-scale spatial feature extraction residual network (MFERN) for hyperspectral image classification. MFERN divides the input data into many non-overlapping sub-bands by spectral bands, extracts features in parallel using the multi-scale spatial feature extraction module MSFE, and adds global branches on top of this to obtain global information of the full spectral band of the image. Finally, the extracted features are fused and sent into the classifier. Our MSFE module has multiple branches with increasing ranges of the receptive field (RF), enabling multi-scale spatial information extraction at both fine- and coarse-grained levels. On the Indian Pines (IP), Salinas (SA), and Pavia University (PU) HSI datasets, we conducted extensive experiments. The experimental results show that our model has the best performance and robustness, and our proposed MFERN significantly outperforms other models in terms of classification accuracy, even with a small amount of training data.

MBT: Model-Based Transformer for retinal optical coherence tomography image and video multi-classification

Background and objectiveThe detection of retinal diseases using optical coherence tomography (OCT) images and videos is a concrete example of a data classification problem. In recent years, Transformer architectures have been successfully applied to solve a variety of real-world classification problems. Although they have shown impressive discriminative abilities compared to other state-of-the-art models, improving their performance is essential, especially in healthcare-related problems. MethodsThis paper presents an effective technique named model-based transformer (MBT). It is based on popular pre-trained transformer models, particularly, vision transformer, swin transformer for OCT image classification, and multiscale vision transformer for OCT video classification. The proposed approach is designed to represent OCT data by taking advantage of an approximate sparse representation technique. Then, it estimates the optimal features, and performs data classification. ResultsThe experiments are carried out using three real-world retinal datasets. The experimental results on OCT image and OCT video datasets show that the proposed method outperforms existing state-of-the-art deep learning approaches in terms of classification accuracy, precision, recall, and f1-score, kappa, AUC-ROC, and AUC-PR. It can also boost the performance of existing transformer models, including Vision transformer and Swin transformer for OCT image classification, and Multiscale Vision Transformers for OCT video classification. ConclusionsThis work presents an approach for the automated detection of retinal diseases. Although deep neural networks have proven great potential in ophthalmology applications, our findings demonstrate for the first time a new way to identify retinal pathologies using OCT videos instead of images. Moreover, our proposal can help researchers enhance the discriminative capacity of a variety of powerful deep learning models presented in published papers. This can be valuable for future directions in medical research and clinical practice.

Skin cancer diagnosis: Leveraging deep hidden features and ensemble classifiers for early detection and classification

ProblemCancer is a deadly disease that requires better diagnostics. Early detection improves skin cancer survival. Due to skin lesion dissimilarity, automated image categorization for skin cancer is problematic, making early diagnosis costly and complicated. ObjectiveThis study proposes a deep hidden feature and ensemble classifier skin cancer detection and classification technique. It handles real-time data streaming prediction and dimensionality issues. MethodsThe research work employs the sand cat swarm optimization with ResNet50 (SCSO-ResNet50) method to separate deep hidden features from known features, ensuring accurate predictions. They utilize an improved harmony search (EHS) approach to optimize features and reduce data dimensionality. Ensemble classifiers, including Naive Bayes, random forest, k-nearest neighbor (k-NN), support vector machine (SVM), and linear regression, are used for early cancer diagnosis. The performance of the proposed methodology is evaluated using the Kaggle skin cancer dataset and the ISIC 2019 dataset, comparing it to state-of-the-art classifiers in terms of accuracy, precision, recall, and F-measure. ResultsIn the second layer of skin cancer classification, the ensemble classifiers, baseline classifier, and CNN outperformed each other, improving prediction accuracy. ConclusionThe proposed methodology improved prediction accuracy compared to state-of-the-art classifiers, as evaluated on benchmark datasets. These findings highlight the potential of the suggested approach in enhancing the early diagnosis of skin cancer.

Dual Mutual Robust Graph Convolutional Network for Weakly Supervised Node Classification in Social Networks of Internet of People

Social networks are a crucial component of the Internet of People (IoP) which is the frontier of the next generation of Internet of Things (IoT). Predicting a large number of unknown node labels with few known labels is one of the challenging problems in social networks analysis. Fortunately, Graph Convolution Network (GCN) and subsequent variants have achieved impressive performance on Semi-Supervised Node Classification (SSNC). However, existing works only focus on the case of clean labels and rarely study the problem of SSNC under Noisy Labels (SSNCNL), which is a more challenging and practical problem in the field of weakly supervised learning. To address this challenge, in this paper, we propose a novel Dual Mutual Robust Graph Convolutional Network named DMRGCN, which is inspired by deep mutual learning and robust learning in the domain of image recognition. Specifically, we first employ two GCNs with different learning abilities to construct network architecture. Then we define a new loss function which is a weighted sum of the supervised loss, mutual loss, and robust loss. Finally, we train the network in the paradigm of pseudo-siamese network. Experimental results on three social network benchmark datasets with different levels of noise on labels demonstrate that DMRGCN outperforms the vanilla GCN and several variants in terms of classification accuracy. In particular, under the two conditions of clean labels and noisy labels, the accuracy of node classification obtained by our proposed DMRGCN can be 3.05% and 6.44% higher than that of the vanilla GCN, respectively.

Confidentiality Preserved Federated Learning for Indoor Localization Using Wi-Fi Fingerprinting

For the establishment of future ubiquitous location-aware applications, a scalable indoor localization technique is essential technology. Numerous classification techniques for indoor localization exist, but none have proven to be as quick, secure, and dependable as what is now needed. This research proposes an effective and privacy-protective federated architecture-based framework for location classification via Wi-Fi fingerprinting. The federated indoor localization classification (f-ILC) system that was suggested had distributed client–server architecture with data privacy for any and all related edge devices or clients. To try and evaluate the proposed f-ILC framework, different data from different sources on the Internet were collected and given in a format that had already been processed. Experiments were conducted with standard learning, federated learning with a single client, and federated learning with several clients to make sure that federated deep learning models worked correctly. The success of the f-ILC framework was computed using a number of factors, such as validation of accuracy and loss. The results showed that the suggested f-ILC framework performed better than traditional distributed deep learning-based classifiers in terms of accuracy and loss while keeping data secure. Due to its innovative design and superior performance over existing classifier tools, edge devices’ data privacy makes this proposed architecture the ideal solution.

Determinants of Inadequate Empiric Antimicrobial Therapy in ICU Sepsis Patients in Al-Madinah Al-Munawwarah, Saudi Arabia: A Comparison of Artificial Neural Network and Regression Analysis.

In the management of sepsis, providing adequate empiric antimicrobial therapy is one of the most important pillars of sepsis management. Therefore, it is important to evaluate the adequacy of empiric antimicrobial therapy (EAMT) in sepsis patients admitted to intensive care units (ICU) and to identify the determinants of inadequate EAMT. The aim of this study was to evaluate the adequacy of empiric antimicrobial therapy in patients admitted to the ICU with sepsis or septic shock, and the determinants of inadequate EAMT. The data of patients admitted to the ICU units due to sepsis or septic shock in two tertiary healthcare facilities in Al-Madinah Al-Munawwarah were retrospectively reviewed. The current study used logistic regression analysis and artificial neural network (ANN) analysis to identify determinants of inadequate empiric antimicrobial therapy, and evaluated the performance of these two approaches in predicting the inadequacy of EAMT. The findings of this study showed that fifty-three per cent of patients received inadequate EAMT. Determinants for inadequate EAMT were APACHE II score, multidrug-resistance organism (MDRO) infections, surgical history (lower limb amputation), and comorbidity (coronary artery disease). ANN performed as well as or better than logistic regression in predicating inadequate EAMT, as the receiver operating characteristic area under the curve (ROC-AUC) of the ANN model was higher when compared with the logistic regression model (LRM): 0.895 vs. 0.854. In addition, the ANN model performed better than LRM in predicting inadequate EAMT in terms of classification accuracy. In addition, ANN analysis revealed that the most important determinants of EAMT adequacy were the APACHE II score and MDRO. In conclusion, more than half of the patients received inadequate EAMT. Determinants of inadequate EAMT were APACHE II score, MDRO infections, comorbidity, and surgical history. This provides valuable inputs to improve the prescription of empiric antimicrobials in Saudi Arabia going forward. In addition, our study demonstrated the potential utility of applying artificial neural network analysis in the prediction of outcomes in healthcare research.

An Unsupervised Feature Extraction Using Endmember Extraction and Clustering Algorithms for Dimension Reduction of Hyperspectral Images

Hyperspectral images (HSIs) provide rich spectral information, facilitating many applications, including landcover classification. However, due to the high dimensionality of HSIs, landcover mapping applications usually suffer from the curse of dimensionality, which degrades the efficiency of supervised classifiers due to insufficient training samples. Feature extraction (FE) is a popular dimension reduction strategy for this issue. This paper proposes an unsupervised FE algorithm that involves extracting endmembers and clustering spectral bands. The proposed method first extracts existing endmembers from the HSI data via a vertex component analysis method. Using these endmembers, it subsequently constructs a prototype space (PS) in which each spectral band is represented by a point. Similar/correlated bands in the PS remain near one another, forming several clusters. Therefore, our method, in the next step, clusters spectral bands into multiple clusters via K-means and fuzzy C-means algorithms. Finally, it combines all the spectral bands in the same cluster using a weighted average operator to decrease the high dimensionality. The extracted features were evaluated by applying an SVM classifier. The experimental results confirmed the superior performance of the proposed method compared with five state-of-the-art dimension reduction algorithms. It outperformed these algorithms in terms of classification accuracy on three widely used hyperspectral images (Indian Pines, KSC, and Pavia Centre). The suggested technique also showed comparable or even stronger performance (up to 9% improvement) compared with its supervised competitor. Notably, the proposed method exhibited higher accuracy even when only a limited number of training samples were available for supervised classification. Using only five training samples per class for the KSC and Pavia Centre datasets, our method’s classification accuracy was higher than that of its best-performing unsupervised competitors by about 7% and 1%, respectively, in our experiments.

Classification of Epileptic and Psychogenic Nonepileptic Seizures via Time-Frequency Features of EEG Data.

The majority of psychogenic nonepileptic seizures (PNESs) are brought on by psychogenic causes, but because their symptoms resemble those of epilepsy, they are frequently misdiagnosed. Although EEG signals are normal in PNES cases, electroencephalography (EEG) recordings alone are not sufficient to identify the illness. Hence, accurate diagnosis and effective treatment depend on long-term video EEG data and a complete patient history. Video EEG setup, however, is more expensive than using standard EEG equipment. To distinguish PNES signals from conventional epileptic seizure (ES) signals, it is crucial to develop methods solely based on EEG recordings. The proposed study presents a technique utilizing short-term EEG data for the classification of inter-PNES, PNES, and ES segments using time-frequency methods such as the Continuous Wavelet transform (CWT), Short-Time Fourier transform (STFT), CWT-based synchrosqueezed transform (WSST), and STFT-based SST (FSST), which provide high-resolution time-frequency representations (TFRs). TFRs of EEG segments are utilized to generate 13 joint TF (J-TF)-based features, four gray-level co-occurrence matrix (GLCM)-based features, and 16 higher-order joint TF moment (HOJ-Mom)-based features. These features are then employed in the classification procedure. Both three-class (inter-PNES versus PNES versus ES: ACC: 80.9%, SEN: 81.8%, and PRE: 84.7%) and two-class (Inter-PNES versus PNES: ACC: 88.2%, SEN: 87.2%, and PRE: 86.1%; PNES versus ES: ACC: 98.5%, SEN: 99.3%, and PRE: 98.9%) classification algorithms performed well, according to the experimental results. The STFT and FSST strategies surpass the CWT and WSST strategies in terms of classification accuracy, sensitivity, and precision. Moreover, the J-TF-based feature sets often perform better than the other two.