Increase Classification Accuracy Research Articles

User-agnostic adaptation of human locomotion intent: Leveraging Teacher-Student-Learning and ensemble modeling

Wearable robot control in complicated contexts requires an understanding of human locomotion intent and behaviors. However, signals from human–robot interfaces are typically reliant on the user, leading to poor performance of the model trained on training user when applied to new end-point users. This study aims to address this problem by developing a novel Teacher-Student-Learning (TSL) approach using Heterogeneous Ensemble Hypotheses (HEH) to achieve unsupervised user-agnostic adaptation. The motivation behind using the Teacher-Student-Learning architecture is to leverage the diverse features extracted by HEH while maintaining computational efficiency. HEH reduces the difference between labeled training users and unlabeled end-point users by incorporating multiple diverse feature generators to extract a wide range of features, thereby increasing classification accuracy and reducing variance. However, this approach sacrifices efficiency due to the ensemble nature of HEH. To address this trade-off, the knowledge from HEH is distilled into a single network using TSL, ensuring precision and efficiency. The proposed approach is evaluated on two publicly available human locomotion datasets and a 2D moon dataset. Experimental results demonstrate the effectiveness of the proposed method, outperforming all other methods. Tested with three datasets, the proposed method can classify end-point users’ data with an average accuracy of 98.9%, 96.7%, and 96.9% while yielding a low processing time (1 ms). Compared to a benchmark method, the suggested strategy improves the average accuracy by 1.5% and 7.2% for categorizing the target users’ locomotor modes and stabilizes the learning curves. The proposed method represents a significant contribution to the field of human activity recognition and human intent prediction for human–robot interaction systems, enhancing the efficiency and generalization capacity of these systems.

Improving golden jackel optimization algorithm: An application of chemical data classification

One of the main issues affecting the effectiveness of the quantitative structure-activity relationship (QSAR) classification techniques in chemometrics is high dimensionality. Applying feature selection is a critical procedure that determines the most relevant and important aspects of a dataset. It improves the effectiveness and accuracy of prediction models by effectively lowering the number of features. This decrease increases classification accuracy, reduces computing strain, and improves overall performance. Recently, the golden jackal optimization (GJO) algorithm was introduced, which has been successfully used to solve various continuous optimization issues. Therefore, this study proposes an improvement in the GJO algorithm employing chaotic maps, abbreviated as CGJO, to enhance the exploration and exploitation capability of the GJO algorithm in picking the essential descriptors in QSAR classification models with high classification accuracy and less computation time. Experimental findings based on four different high-dimensional chemical datasets show that the proposed CGJO algorithm can maximize classification accuracy while simultaneously decreasing the number of chosen descriptors and lowering the time required for computing. Thus, the proposed algorithm can be useful for chemical data classification in other QSAR modeling.

Motion Signal-based Recognition of Human Activity from Video Stream Dataset Using Deep Learning Approach

Background: Human physical activity recognition is challenging in various research eras, such as healthcare, surveillance, senior monitoring, athletics, and rehabilitation. The use of various sensors has attracted outstanding research attention due to the implementation of machine learning and deep learning approaches. Aim: This paper proposes a unique deep learning framework based on motion signals to recognize human activity to handle these constraints and challenges through deep learning (e.g., Enhance CNN, LR, RF, DT, KNN, and SVM) approaches. Method: This research article uses the BML (Biological Motion Library) dataset gathered from thirty volunteers with four various activities to analyze the performance metrics. It compares the evaluated results with existing results, which are found by machine learning and deep learning methods to identify human activity. Result: This framework was successfully investigated with the help of laboratory metrics with convolutional neural networks (CNN) and achieved 89.0% accuracy compared to machine learning methods. Conclusion: The novel work of this research is to increase classification accuracy with a lower error rate and faster execution. Moreover, it introduces a novel approach to human activity recognition in the BML dataset using the CNN with Adam optimizer approach.

Plant Leaf Disease Detection

Abstract: The research presents an automated vision a system that makes use of the processing of images techniques to identify plant diseases in agricultural contexts. In order to monitor vast crop fields and automatically identify disease symptoms as soon that they occur on plant leaves, research on automated identification of plant infections is crucial to the agricultural industry. This method uses segmentation, colour modifications, and masking of green pixels to classify data based on learning from some training examples of that category. The simulated outcome, in the end, demonstrates that the network classifier in use offers reduced training error and increased classification accuracy

Detection of Sedge Weeds Infestation in Wetland Rice Cultivation Using Hyperspectral Images and Artificial Intelligence: A Review

Sedge is one type of weed that can infest the rice field, as well as broadleaf and grasses. If sedges are not appropriately controlled, severe yield loss will occur due to increased competition with cultivated rice for light, space, nutrients, and water. Both sedges and grasses are monocots and have similar narrowed leaf characteristics, but most sedge stems have triangular prismatic shapes in cross sections, which differ them from grasses. Event sedges and grasses differ in morphology, but differentiating them in rice fields is challenging due to the large rice field area and high green color similarity. In addition, climate change makes it more challenging as the distribution of sedge weed infestation is influenced by surrounding abiotic factors, which lead to changes in weed control management. With advanced drone technology, agriculture officers or scientists can save time and labor in distributing weed surveys in rice fields. Using hyperspectral sensors on drones can increase classification accuracy and differentiation between weed species. The spectral signature of sedge weed species captured by the hyperspectral drone can generate weed maps in rice fields to give the sedge percentage distribution and location of sedge patch growth. Researchers can propose proper countermeasures to control the sedge weed problem with this information. This review summarizes the advances in our understanding of the hyperspectral reflectance of weedy sedges in rice fields. It also discusses how they interact with climate change and phenological stages to predict sedge invasions.

A Comparative Analysis of Combination of CNN-Based Models with Ensemble Learning on Imbalanced Data

This study investigates the usefulness of the Synthetic Minority Oversampling Technique (SMOTE) in conjunction with convolutional neural network (CNN) models, which include both single and ensemble classifiers. The objective of this research is to handle the difficulty of multi-class imbalanced image classification. The application of SMOTE in imbalanced picture datasets is still underexplored, even though CNNs have been shown to be successful in image classification and that ensemble learning approaches have improved their performance. To investigate whether or not SMOTE can increase classification accuracy and other performance measures when combined with CNN-based classifiers, our research makes use of a CIFAR-10 dataset that has been artificially step-imbalanced and has varying imbalanced ratios. We conducted experiments using five distinct models, namely AdaBoost, XGBoost, standalone CNN, CNN-AdaBoost, and CNN-XGBoost, on datasets that were either imbalanced or SMOTE-balanced. Metrics such as accuracy, precision, recall, F1-score, and the area under the receiver operating characteristic curve (AUC) were included in the evaluation process. The findings indicate that SMOTE dramatically improves the accuracy of minority classes, and that the combination of ensemble classifiers with CNNs and oversampling techniques significantly improves overall classification performance, particularly in situations when there is a high-class imbalance. When it comes to enhancing imbalanced classification tasks, this study demonstrates the potential of merging oversampling techniques with CNN-based ensemble classifiers to minimize the impacts of class imbalance in picture datasets. This suggests a promising direction for future research in this area.

Forecasting Epileptic Seizures Using XGBoost Methodology and EEG Signals

INTRODUCTION: Epilepsy denotes a disorder of neurological origin marked by repetitive and spontaneous seizures without any apparent trigger. Seizures occur due to abrupt and heightened electricity flowing through the brain, which can lead to physical and mental symptoms. There are several types of epileptic seizures, and epilepsy itself can be caused by various underlying conditions. EEG (Electroencephalogram) is one of the most important and widely used tools for epileptic seizure prediction and diagnosis. EEG uses skull sensors to record electrical signals from the brain., and it can provide valuable insights into brain activity patterns associated with seizures.
 OBJECTIVES: Brain-computer interface technology pathway for analyzing the EEG signals for seizure prediction to eliminate the class imbalance issue from our dataset in this case, a SMOTE approach is applied. It is observable that there are more classes of one variable than there are of the others in the output variable. This will be problematic when employing different Artificial intelligence techniques since these algorithms are more likely to be biased towards a certain variable because of its high prevalence
 METHODS: SMOTE approaches will be used to address this bias and balance the number of variables in the response variable. To develop an XGBoost (Extreme Gradient Boosting) model using SMOTE techniques to increase classification accuracy.
 RESULTS: The results show that the XGBoost method achieves a 98.7% accuracy rate.
 CONCLUSION: EEG-based model for seizure type using the XGBoost model for predicting the disease early. The Suggested method could significantly reduce the amount of time needed to accomplish seizure prediction.

Enhancing Transparency and Interpretability in Toxic Comment Classification: A Study on the Integration of Explainable Artificial Intelligence (XAI) Techniques || Dr. Pallavi Devendra Tawde, Mr. Jadyn Dias

More than ever, robust, and interpretable toxic comment recognition methods are required to manage the growing frequency of toxic comments on online platforms. The research tries to incorporate techniques in Explainable Artificial Intelligence (XAI) to improve the transparency and comprehensibility of toxic comment classification. Using a comprehensive dataset, we designed a model architecture which includes the latest practices in XAI. Through rigorous experimentation, our study proves the usefulness of such methods as tools that not only increase classification accuracy but also illuminate model decision-making processes. One view is that by adding LIME and Eli5 to toxic comment classification, model performance improves both in terms of accuracy and interpretation for decisions. Our results provide valuable insights into the model's strengths and areas for refinement, contributing to the transparency and interpretability of toxic comment classification. This research contributes to the evolving landscape of interpretable machine learning, offering a pathway to more accountable and trustworthy toxic comment moderation systems. Keywords: Explainable artificial intelligence, Model interpretability, toxic comment classification, LIME, Eli5

The classification of SSVEP-BCI based on ear-EEG via RandOm Convolutional KErnel Transform with Morlet wavelet

During the advantages of shorter training and higher information throughput, steady-state visual evoked potential (SSVEP) is widely used in brain–computer interface (BCI) research. Recently, collecting EEG signals from the ear area (ear-EEG) has gained increasing attention because it is more comfortable and convenient than scalp electrodes. The ear-EEG-based BCI system based on ear electrodes has weaker signals and more noise components because the electrodes are located far away from the top of the head. In this study, the RandOm Convolutional KErnel Transform (ROCKET) algorithm integrated with the Morlet wavelet transform (Morlet-ROCKET) was proposed to solve this issue. This study compared the performence of Morlet-ROCKET with two established methods: canonical correlation analysis-based (FBCCA) and Transformer methods. The proposed Morlet-ROCKET model demonstrated superior performance across multiple measures, including increased classification accuracy in 1 s, 3 s, and 4 s time windows and higher area under the curve (AUC) values in receiver operating characteristic (ROC) analysis. The analysis result proved that with efficient data processing algorithms, ear-EEG-based BCI systems can also have good performance, and providing support for the popularization of BCI.

Reviewing various feature selection techniques in machine learning‐based botnet detection

SummaryMachine learning approaches are widely used for the detection and classification of emerging botnet variations due to their ability to yield more precise results compared to traditional methods. The relevancy of the features plays a major role in these detection algorithms' effectiveness. As such, the most distinctive characteristics must be extracted from a high‐dimensional dataset that is used to classify botnets. Nevertheless, we discovered that the majority of earlier studies lacked proper analysis and paid little attention to the various feature selection techniques. The main goal of this work is to investigate and assess the advantages and disadvantages of the different feature selection techniques used for botnet detection. Studies show that feature selection is a very efficient way to decrease the amount of storage and processing power required while simultaneously increasing classification accuracy. As a consequence, its application in many other fields has grown. The field of feature selection is recognized for its non‐deterministic polynomial‐time hardness; to mitigate this hardness, metaheuristic techniques have been applied. Metaheuristic algorithms are exceptionally good at performing a global search. In order to choose feature subsets optimally in the field of botnet detection, we additionally prioritize the use of metaheuristic methods. This study offers a more thorough insight of the feature selection strategies that are primarily employed by machine learning‐based botnet detection models. It also offers insights into how better feature selection approaches might be applied to strengthen botnet detection mechanisms. Additionally, it will help in understanding the limitations of existing approaches and identifying areas for improvement.

Early prediction of mastery of a computerized functional skills training program in participants with mild cognitive impairment.

Cognition in MCI has responded poorly to pharmacological interventions, leading to use of computerized training. Combining computerized cognitive training (CCT) and functional skills training software (FUNSAT) produced improvements in 6 functional skills in MCI, with effect sizes >0.75. However, 4% of HC and 35% of MCI participants failed to master all 6 tasks. We address early identification of characteristics that identify participants who do not graduate, to improve later interventions. NC participants (n = 72) received FUNSAT and MCI (n = 92) participants received FUNSAT alone or combined FUNSAT and CCT on a fully remote basis. Participants trained twice a week for up to 12 weeks. Participants "graduated" each task when they made one or fewer errors on all 3-6 subtasks per task. Tasks were no longer trained after graduation. Between-group comparisons of graduation status on baseline completion time and errors found that failure to graduate was associated with more baseline errors on all tasks but no longer completion times. A discriminant analysis found that errors on the first task (Ticket purchase) uniquely separated the groups, F = 41.40, p < .001, correctly classifying 94% of graduators. An ROC analysis found an AUC of .83. MOCA scores did not increase classification accuracy. More baseline errors, but not completion times, predicted failure to master all FUNSAT tasks. Accuracy of identification of eventual mastery was exceptional. Detection of risk to fail to master training tasks is possible in the first 15 minutes of the baseline assessment. This information can guide future enhancements of computerized training.

Pattern recognition frequency-based feature selection with multi-objective discrete evolution strategy for high-dimensional medical datasets

Feature selection has a prominent role in high-dimensional datasets to increase classification accuracy, decrease the learning algorithm computational time, and present the most informative features to decision-makers. This paper proposes a two-stage hybrid feature selection for high-dimensional medical datasets: Maximum Pattern Recognition - Multi-objective Discrete Evolution Strategy (MPR-MDES). MPR is a rapid filter ranker that significantly outperforms existing frequency-based rankers in recognizing non-linear patterns, effectively eliminating a majority of non-informative features. Then, the wrapper Multi-objective Discrete Evolution Strategy (MDES) uses the remaining features and obtains sets of solutions which are automatically presented to decision-makers. The experiments conducted on large medical datasets demonstrate that MPR-MDES achieves considerable improvements compared to state-of-the-art methods, in terms of both classification accuracy and dimensionality reduction. In this sense, the proposal successfully performs when presenting informative feature sets to decision-makers. The implementation is available on https://github.com/KhaosResearch/MPR-MDES.

Enabling uncertainty estimation in neural networks through weight perturbation for improved Alzheimer's disease classification.

The willingness to trust predictions formulated by automatic algorithms is key in a wide range of domains. However, a vast number of deep architectures are only able to formulate predictions without associated uncertainty. In this study, we propose a method to convert a standard neural network into a Bayesian neural network and estimate the variability of predictions by sampling different networks similar to the original one at each forward pass. We combine our method with a tunable rejection-based approach that employs only the fraction of the data, i.e., the share that the model can classify with an uncertainty below a user-set threshold. We test our model in a large cohort of brain images from patients with Alzheimer's disease and healthy controls, discriminating the former and latter classes based on morphometric images exclusively. We demonstrate how combining estimated uncertainty with a rejection-based approach increases classification accuracy from 0.86 to 0.95 while retaining 75% of the test set. In addition, the model can select the cases to be recommended for, e.g., expert human evaluation due to excessive uncertainty. Importantly, our framework circumvents additional workload during the training phase by using our network "turned into Bayesian" to implicitly investigate the loss landscape in the neighborhood of each test sample in order to determine the reliability of the predictions. We believe that being able to estimate the uncertainty of a prediction, along with tools that can modulate the behavior of the network to a degree of confidence that the user is informed about (and comfortable with), can represent a crucial step in the direction of user compliance and easier integration of deep learning tools into everyday tasks currently performed by human operators.

Multimodal and Multidomain Feature Fusion for Emotion Classification Based on Electrocardiogram and Galvanic Skin Response Signals

Emotion classification using physiological signals is a promising approach that is likely to become the most prevalent method. Bio-signals such as those derived from Electrocardiograms (ECGs) and the Galvanic Skin Response (GSR) are more reliable than facial and voice recognition signals because they are not influenced by the participant’s subjective perception. However, the precision of emotion classification with ECG and GSR signals is not satisfactory, and new methods need to be developed to improve it. In addition, the fusion of the time and frequency features of ECG and GSR signals should be explored to increase classification accuracy. Therefore, we propose a novel technique for emotion classification that exploits the early fusion of ECG and GSR features extracted from data in the AMIGOS database. To validate the performance of the model, we used various machine learning classifiers, such as Support Vector Machine (SVM), Decision Tree, Random Forest (RF), and K-Nearest Neighbor (KNN) classifiers. The KNN classifier gives the highest accuracy for Valence and Arousal, with 69% and 70% for ECG and 96% and 94% for GSR, respectively. The mutual information technique of feature selection and KNN for classification outperformed the performance of other classifiers. Interestingly, the classification accuracy for the GSR was higher than for the ECG, indicating that the GSR is the preferred modality for emotion detection. Moreover, the fusion of features significantly enhances the accuracy of classification in comparison to the ECG. Overall, our findings demonstrate that the proposed model based on the multiple modalities is suitable for classifying emotions.

Optimal Image Characterization for In-Bed Posture Classification by Using SVM Algorithm

Identifying patient posture while they are lying in bed is an important task in medical applications such as monitoring a patient after a surgical intervention, sleep supervision to identify behavioral and physiological markers, or for bedsore prevention. An acceptable strategy to identify the patient’s position is the classification of images created from a grid of pressure sensors located in the bed. These samples can be arranged based on supervised learning methods. Usually, image conditioning is required before images are loaded into a learning method to increase classification accuracy. However, continuous monitoring of a person requires large amounts of time and computational resources if complex pre-processing algorithms are used. So, the problem is to classify the image posture of patients with different weights, heights, and positions by using minimal sample conditioning for a specific supervised learning method. In this work, it is proposed to identify the patient posture from pressure sensor images by using well-known and simple conditioning techniques and selecting the optimal texture descriptors for the Support Vector Machine (SVM) method. This is in order to obtain the best classification and to avoid image over-processing in the conditioning stage for the SVM. The experimental stages are performed with the color models Red, Green, and Blue (RGB) and Hue, Saturation, and Value (HSV). The results show an increase in accuracy from 86.9% to 92.9% and in kappa value from 0.825 to 0.904 using image conditioning with histogram equalization and a median filter, respectively.

Fault diagnosis using signal processing and deep learning-based image pattern recognition

Abstract The vibration signal is a typical non-stationary signal, making it challenging to use traditional time-frequency analysis techniques for fault diagnosis. Therefore, this work investigates the processing of vibration signals and proposes a deep learning method based on processed signals for the fault diagnosis of ball bearings. In this work, the fault diagnosis is formulated as an image classification problem and solved with deep learning networks. The intrinsic mode functions (IMFs), converted from the vibration signals in the time domain, are then transformed into symmetrized dot pattern (SDP) images. In order to increase classification accuracy, the SDP parameters in this study are chosen by optimizing image similarity. The feasibility and accuracy of the proposed approach are examined experimentally.

A deep learning fusion model for accurate classification of brain tumours in Magnetic Resonance images

AbstractDetecting brain tumours is complex due to the natural variation in their location, shape, and intensity in images. While having accurate detection and segmentation of brain tumours would be beneficial, current methods still need to solve this problem despite the numerous available approaches. Precise analysis of Magnetic Resonance Imaging (MRI) is crucial for detecting, segmenting, and classifying brain tumours in medical diagnostics. Magnetic Resonance Imaging is a vital component in medical diagnosis, and it requires precise, efficient, careful, efficient, and reliable image analysis techniques. The authors developed a Deep Learning (DL) fusion model to classify brain tumours reliably. Deep Learning models require large amounts of training data to achieve good results, so the researchers utilised data augmentation techniques to increase the dataset size for training models. VGG16, ResNet50, and convolutional deep belief networks networks extracted deep features from MRI images. Softmax was used as the classifier, and the training set was supplemented with intentionally created MRI images of brain tumours in addition to the genuine ones. The features of two DL models were combined in the proposed model to generate a fusion model, which significantly increased classification accuracy. An openly accessible dataset from the internet was used to test the model's performance, and the experimental results showed that the proposed fusion model achieved a classification accuracy of 98.98%. Finally, the results were compared with existing methods, and the proposed model outperformed them significantly.

DEEP NETWORK-BASED METHOD AND SOFTWARE FOR SMALL SAMPLE BIOMEDICAL IMAGE GENERATION AND CLASSIFICATION

Context. The authors of the article investigated the problem of generating and classifying breast cancer histological images. The widespread incidence of breast cancer explains the problem’s relevance. The automated diagnosing procedure saves time and eliminates the subjective aspect. The study’s findings can be applied to cancer CAD systems.&#x0D; Objective. The purpose of the study is to develop a deep neural network-based method and software tool for generating and classifying histological images in order to increase classification accuracy.&#x0D; Method. The method of histological image generation and classification was developed in the research study. This method employs CNN and GAN. To improve the classification accuracy, the initial image sample was expanded using GAN.&#x0D; Results. The computer research of the developed method of image generation and classification was conducted on the basis of the dataset located on the Zenodo platform. Light microscopy served as the basis for obtaining the image. The dataset contained three classes of G1, G2, and G3 breast cancer histological images. Based on the developed method, the accuracy of image classification was 96%. This is a higher classification accuracy compared to existing models such as AlexNet, LeNet5, and VGG16. The software module can be integrated into CAD.&#x0D; Conclusions. The developed method of generating and classifying images is the basis of the software module. The software module can be integrated into CAD.

Evolution of the random subset feature selection algorithm for classification problem

Datasets often include excessive or irrelevant data that affect the performance and complexity of the machine learning model. Feature selection is one of the most effective dimension-reduction technics in datasets to solve the problem. Feature reduction can lead to increased classification accuracy and decreased computational costs. Wrapper feature selection algorithms are a popular and effective category of these methods which considering the learning method's feedback. One of the well-known algorithms in this area is the random subset feature selection algorithm (RSFS). This study proposes an improved version that has higher convergence speed, lower feature selection rate, and higher classification accuracy. Improvements include considering the features' interaction in the selective subsets, continuous evaluation of the selected features to avoid being stuck in the local optima and enhancing the method's evolution phase. To comprehensively evaluate the proposed algorithm, its performance was applied to 20 standard and well-known datasets from public resources and compared with various recent related methods. Experimental results demonstrate that the proposed method reduced classification errors using fewer features compared to other methods, achieving the highest ranking in the Friedman and Wilcoxon rank-sum tests.

Improved discrimination of COVID-19 based on data enhancement technology and an information balance feature selection (INB) method

The coronavirus disease (COVID-19) ravaged the world in late 2019 and posed a serious threat to human life and property destruction on a global scale. In this paper, the Wasserstein generative adversarial network with gradient penalty (WGAN-GP) method was selected for balancing the data sample, and an information balance feature selection (INB) method was first proposed to realize the accurate discrimination of COVID-19 saliva samples based on the attenuated total reflection Fourier transform infrared (ATR-FTIR) spectroscopy. The results of the experiment showed that the INB method obtained higher classification accuracy than the traditional feature selection methods in both the original spectrum and the second-order derivative spectrum, especially in the second-order derivative spectrum where all the indexes reached about 85 %. In addition, the combination of WGAN_GP data augmentation and the INB method resulted in an accuracy of 88.7 % for the original spectrum and even 90.6 % for the second-order derivative spectrum. According to these findings, classification research using the WGAN_GP data enhancement model may increase classification accuracy. Additionally, the ability to successfully separate COVID-19 indicates that the INB method to identify spectral data features is a workable method, which also offers a fresh viewpoint on feature selection.