Redundancy Maximum Relevance Feature Selection Research Articles

Energy optimization for wireless sensor network using minimum redundancy maximum relevance feature selection and classification techniques

In wireless sensor networks (WSN), conserving energy is usually a basic issue, and several approaches are applied to optimize energy consumption. In this article, we adopt feature selection approaches by using minimum redundancy maximum relevance (MRMR) as a feature selection technique to minimize the number of sensors thereby conserving energy. MRMR ranks the sensors according to their significance. The selected features are then classified by different types of classifiers; SVM with linear kernel classifier, naïve Bayes classifier, and k-nearest neighbors classifier (KNN) to compare accuracy values. The simulation results illustrated an improvement in the lifetime extension factor of sensors and showed that the KNN classifier gives better results than the naïve Bayes and SVM classifier.

Fed-mRMR: A lossless federated feature selection method

Feature selection has become a mandatory task in data mining, due to the overwhelming amount of features in Big Data problems. To handle this high-dimensional data and avoid the well-known curse of dimensionality, we need to pre-select an optimal subset of features to reduce redundant computations. Federated learning is a machine learning technique based on training an algorithm over many decentralized edge devices holding local rather than global data on a centralized server. Application of this technique is extending to fields such as self-driving cars, medicine and health, and Industry 4.0, where data privacy is compulsory. Feature selection through federated learning is a complicated task since suboptimal features calculated by feature selection methods may be different in heterogeneous datasets from different nodes. In this paper, we propose a lossless federated version of the classic minimum redundancy maximum relevance (mRMR) feature selection algorithm, called federated mRMR (fed-mRMR), which, without losing any effectiveness of the original mRMR method, is applicable to federated learning approaches and capable of dealing with data that are not independent and identically distributed (non-IID data).Implementation can be found at: https://github.com/jorgehermo9/fed-mrmr

Predictive modeling for breast cancer based on machine learning algorithms and features selection methods

Breast cancer is one of the leading causes of death among women worldwide. However, early prediction of breast cancer plays a crucial role. Therefore, strong needs exist for automatic accurate early prediction of breast cancer. In this paper, machine learning (ML) classifiers combined with features selection methods are used to build an intelligent tool for breast cancer prediction. The Wisconsin diagnostic breast cancer (WDBC) dataset is used to train and test the model. Classification algorithms, including support vector machine (SVM), light gradient boosting machine (LightGBM), random forest (RF), logistic regression (LR), k-nearest neighbors (k-NN), and naïve Bayes, were employed. Performance measures for each of them were obtained, namely: accuracy, precision, recall, F-score, Kappa, Matthews correlation coefficient (MCC), and time. The results indicate that without feature selection, LightGBM achieves the highest accuracy at 95%. With minimum redundancy maximum relevance (mRMR) feature selection (15 features), LightGBM outperforms other classifiers, achieving an accuracy of 98%. For Pearson correlation coefficient feature selection (15 features), LightGBM also excels with a 95% accuracy rate. Lasso feature selection (5 features) produces varied results across classifiers, with logistic regression achieving the highest accuracy at 96%. These findings underscore the importance of feature selection in refining model performance and in improving detection for breast cancer.

Data-driven learning-based classification model for mitigating false data injection attacks on dynamic line rating systems

The increasing need to explore electric power grid expansion technologies like dynamic line rating (DLR) systems and their dependence on real-time weather data for system planning necessitates research into false data injection attacks (FDIA) on cyber-physical power systems (CPS). This study aims to develop a robust machine learning model to mitigate FDIA in DLR systems, focusing on statistical data processing, feature ranking, selection, training, validation and evaluation. It synthesises z-score and other statistical analyses with minimum redundancy maximum relevance (MR-MR) feature ranking and selection algorithm to improve model performance and generalisation of binary generalised linear model logistic regression (BGLM-LR) and other machine learning classification algorithms. The resulting models formed with BGLM-LR, Gaussian naïve Bayes (GNB), linear support vector machine (LSVM), wide neural network (WNN), and decision tree (DT) were trained and tested with 10-year hourly DLR history data features. The evaluation of the models on unseen data revealed enhanced validation and testing accuracies after the MR-MR feature ranking and selection. BGLM-LR, GNB, LSVM, and WNN showed promising performance for mitigating FDIA. However, the study identifies DT’s limitations as overfitting and lacking generalisation in FDIA mitigation. z-score-MR-MR-BGLM-LR and z-score-MR-MR-LSVM models exhibited outstanding performances with zero false negative rates highlighting the significance of feature ranking and selection. Still, the z-score-MR-MR-BGLM-LR combination exhibits the highest marginal improvement from training to testing, the lowest training and validation time and a perfect area under curve (AUC) of the receiver operating characteristics making it the best choice in mitigating FDIA when computational resources are limited.

Analysis of selected deep features with CNN-SVM-based for bread wheat seed classification

The main ingredient of flour is processed wheat. Wheat is an agricultural product that is harvested once a year. It may be necessary to choose the variety of wheat for growing wheat and efficient harvesting. The variety of wheat is important for its economic value, taste, and crop yield. Although there are many varieties of wheat, they are very similar in colour, size, and shape, and it requires expertise to distinguish them by eye. This is very time consuming and can lead to human error. Using computer vision and artificial intelligence, such problems can be solved more quickly and objectively. In this study, an attempt was made to classify five bread wheat varieties belonging to different cultivars using Convolutional Neural Network (CNN) models. Three approaches have been proposed for classification. First, pre-trained CNN models (ResNet18, ResNet50, and ResNet101) were trained for bread wheat cultivars. Second, the features extracted from the fc1000 layer of the pre-trained CNN models ResNet18, ResNet50, and ResNet101 were classified using a support vector machine (SVM) classifier with different kernel features from machine learning techniques for classification with different variants. Finally, SVM methods were used in the second stage to classify the features obtained from the fc1000 layer of the pre-trained CNN models with an optimal set of features that can represent all features using the minimum redundancy maximum relevance (mRMR) feature selection algorithm.The accuracies obtained in the first, second, and last phases are as follows. In the first phase, the most successful method in classifying wheat grains was the ResNet18 model with 97.57%. In the second phase, the ResNet18 + ResNet50 + ResNet101 + Quadratic SVM model was the most successful model in classification using the features obtained from the ResNet CNN models with 94.08%.The accuracy for classification with the 1000 most effective features selected by the feature selection algorithm was 94.51%. Although the classification with features is slightly lower than deep learning, the classification time is much shorter and is 93%. This result confirms the great effectiveness of CNN models for wheat grain classification.

Accurate Liver Fibrosis Detection Through Hybrid MRMR-BiLSTM-CNN Architecture with Histogram Equalization and Optimization.

The early detection and accurate diagnosis of liver fibrosis, a progressive and potentially serious liver condition, are crucial for effective medical intervention. Invasive methods like biopsies for diagnosis can be risky and expensive. This research presents a novel computer-aided diagnosis model for liver fibrosis using a hybrid approach of minimum redundancy maximum relevance (MRMR) feature selection, bidirectional long short-term memory (BiLSTM), and convolutional neural networks (CNN). The proposed model involves multiple stages, including image acquisition, preprocessing, feature representation, fibrous tissue identification, and classification. Notably, histogram equalization is employed to enhance image quality by addressing variations in brightness levels. Performance evaluation encompasses a range of metrics such as accuracy, precision, sensitivity, specificity, F1 score, and error rate. Comparative analyses with established methods like DCNN, ANN-FLI, LungNet22, and SDAE-GAN underscore the efficacy of the proposed model. The innovative integration of hybrid MRMR-BiLSTM-CNN architecture and the horse herd optimization algorithm significantly enhances accuracy and F1 score, even with small datasets. The model tackles the complexities of hyperparameter optimization through the IHO algorithm and reduces training time by leveraging MRMR feature selection. In practical application, the proposed hybrid MRMR-BiLSTM-CNN method demonstrates remarkable performance with a 97.8% accuracy rate in identifying liver fibrosis images. It exhibits high precision, sensitivity, specificity, and minimal error rate, showcasing its potential for accurate and non-invasive diagnosis.

SEMG-based Sarcopenia risk classification using empirical mode decomposition and machine learning algorithms.

Early detection of the risk of sarcopenia at younger ages is crucial for implementing preventive strategies, fostering healthy muscle development, and minimizing the negative impact of sarcopenia on health and aging. In this study, we propose a novel sarcopenia risk detection technique that combines surface electromyography (sEMG) signals and empirical mode decomposition (EMD) with machine learning algorithms. First, we recorded and preprocessed sEMG data from both healthy and at-risk individuals during various physical activities, including normal walking, fast walking, performing a standard squat, and performing a wide squat. Next, electromyography (EMG) features were extracted from a normalized EMG and its intrinsic mode functions (IMFs) were obtained through EMD. Subsequently, a minimum redundancy maximum relevance (mRMR) feature selection method was employed to identify the most influential subset of features. Finally, the performances of state-of-the-art machine learning (ML) classifiers were evaluated using a leave-one-subject-out cross-validation technique, and the effectiveness of the classifiers for sarcopenia risk classification was assessed through various performance metrics. The proposed method shows a high accuracy, with accuracy rates of 0.88 for normal walking, 0.89 for fast walking, 0.81 for a standard squat, and 0.80 for a wide squat, providing reliable identification of sarcopenia risk during physical activities. Beyond early sarcopenia risk detection, this sEMG-EMD-ML system offers practical values for assessing muscle function, muscle health monitoring, and managing muscle quality for an improved daily life and well-being.

EEG-Based Depression Detection in the Prefrontal Cortex Lobe using mRMR Feature Selection and Bidirectional LSTM

Depression can induce significant anguish and impair one's ability to perform effectively in professional, academic, and familial settings. This condition has the potential to result in suicide. Annually, the number of deaths resulting from suicide exceeds 700,000. Among individuals aged 15-29, suicide has emerged as the fourth most prevalent cause of mortality. Challenges in treating depression include limited accessibility to mental health care in rural regions and misdiagnosis resulting from subjective evaluations, wherein insufficient expertise can contribute to inaccurate diagnoses. Electroencephalography (EEG) has gained popularity as a tool for the identification and study of a number of mental illnesses in the past several years. Therefore, an automated technique is required to precisely distinguish between normal EEG signals and depression signals. This research focused on developing an EEG-based depression detection system in the prefrontal cortex lobe area (Fp1, Fpz, and Fp2). One of the developments carried out in this research is the implementation of Bidirectional Long Short-Term Memory (Bi-LSTM) as the model classification and minimum redundancy maximum relevance (mRMR) feature selection. Results suggest that the combination of mRMR feature selection with 25 features and the Bidirectional LSTM obtained 92.83% for the accuracy.

Application of Machine Learning Strategies to Model the Effects of Sevoflurane on Somatosensory-Evoked Potentials during Spine Surgery

In this study, a small sample of patients’ neuromonitoring data was analyzed using machine learning (ML) tools to provide proof of concept for quantifying complex signals. Intraoperative neurophysiological monitoring (IONM) is a valuable asset for monitoring the neurological status of a patient during spine surgery. Notably, this technology, when operated by neurophysiologists and surgeons familiar with proper alarm criteria, is capable of detecting neurological deficits. However, non-surgical factors, such as volatile anesthetics like sevoflurane, can negatively influence robust IONM signal generation. While sevoflurane has been shown to affect the latency and amplitude of somatosensory evoked potential (SSEP), a more complex and nuanced analysis of the SSEP waveform has not been performed. In this study, signal processing and machine learning techniques were used to more intricately characterize and predict SSEP waveform changes as a function of varying end-tidal sevoflurane concentration. With data from ten patients who underwent spinal procedures, features describing the SSEP waveforms were generated using principal component analysis (PCA), phase space curves (PSC), and time-frequency analysis (TFA). A minimum redundancy maximum relevance (MRMR) feature selection technique was then used to identify the most important SSEP features associated with changing sevoflurane concentrations. Once the features carrying the maximum amount of information about the majority of signal waveform variability were identified, ML models were used to predict future changes in SSEP waveforms. Linear regression, regression trees, support vector machines, and neural network ML models were then selected for testing. Using SSEP data from eight patients, the models were trained using a range of features selected during MRMR calculations. During the training phase of model development, the highest performing models were identified as support vector machines and regression trees. After identifying the highest performing models for each nerve group, we tested these models using the remaining two patients’ data. We compared the models’ performance metrics using the root mean square error values (RMSEs). The feasibility of the methodology described provides a general framework for the applications of machine learning strategies to further delineate the effects of surgical and non-surgical factors affecting IONM signals.

Identifying Genetic Signatures from Single-Cell RNA Sequencing Data by Matrix Imputation and Reduced Set Gene Clustering

In this current era, the identification of both known and novel cell types, the representation of cells, predicting cell fates, classifying various tumor types, and studying heterogeneity in various cells are the key areas of interest in the analysis of single-cell RNA sequencing (scRNA-seq) data. Due to the nature of the data, cluster identification in single-cell sequencing data with high dimensions presents several difficulties. In this paper, we introduce a new framework that combines various strategies such as imputed matrix, minimum redundancy maximum relevance (MRMR) feature selection, and shrinkage clustering to discover gene signatures from scRNA-seq data. Firstly, we conducted the pre-filtering of the “drop-out” value in the data focusing solely on imputing the identified “drop-out” values. Next, we applied the MRMR feature selection method to the imputed data and obtained the top 100 features based on the MRMR feature selection optimization scores for further downstream analysis. Thereafter, we employed shrinkage clustering on the selected feature matrix to identify the cell clusters using a global optimization approach. Finally, we applied the Limma-Voom R tool employing voom normalization and an empirical Bayes test to detect differentially expressed features with a false discovery rate (FDR) < 0.001. In addition, we performed the KEGG pathway and gene ontology enrichment analysis of the identified biomarkers using David 6.8 software. Furthermore, we conducted miRNA target detection for the top gene markers and performed miRNA target gene interaction network analysis using the Cytoscape online tool. Subsequently, we compared our detected 100 markers with our previously detected top 100 cluster-specified markers ranked by FDR of the latest published article and discovered three common markers; namely, Cyp2b10, Mt1, Alpi, along with 97 novel markers. In addition, the Gene Set Enrichment Analysis (GSEA) of both marker sets also yields similar outcomes. Apart from this, we performed another comparative study with another published method, demonstrating that our model detects more significant markers than that model. To assess the efficiency of our framework, we apply it to another dataset and identify 20 strongly significant up-regulated markers. Additionally, we perform a comparative study of different imputation methods and include an ablation study to prove that every key phase of our framework is essential and strongly recommended. In summary, our proposed integrated framework efficiently discovers differentially expressed stronger gene signatures as well as up-regulated markers in single-cell RNA sequencing data.

Novel Fractal-Based Sub-RPE Compartment OCT Radiomics Biomarkers Are Associated With Subfoveal Geographic Atrophy in Dry AMD.

The purpose of this study was to quantitatively characterize the shape of the sub-retinal pigment epithelium (sub-RPE, i.e., space bounded by RPE and Bruch's membrane) compartment on SD-OCT using fractal dimension (FD) features and evaluate their impact on risk of subfoveal geographic atrophy (sfGA) progression. This was an IRB-approved retrospective study of 137 subjects with dry age-related macular degeneration (AMD) with subfoveal GA. Based on sfGA status at year five, eyes were categorized as "Progressors" and "Non-progressors". FD analysis allows quantification of the degree of shape complexity and architectural disorder associated with a structure. To characterize the structural irregularities along the sub-RPE surface between the two groups of patients, a total of 15 shape descriptors of FD were extracted from the sub-RPE compartment of baseline OCT scans. The top four features were identified using minimum Redundancy maximum Relevance (mRmR) feature selection method and evaluated with Random Forest (RF) classifier using three-fold cross validation from the training set (N = 90). Classifier performance was subsequently validated on the independent test set (N = 47). Using the top four FD features, a RF classifier yielded an AUC of 0.85 on the independent test set. Mean fractal entropy (p-value = 4.8e-05) was identified as the most significant biomarker; higher values of entropy being associated with greater shape disorder and risk for sfGA progression. FD assessment holds promise for identifying high-risk eyes for GA progression. With further validation, FD features could be potentially used for clinical trial enrichment and assessments for therapeutic response in dry AMD patients.

Radiomics predictive modeling from dual-time-point FDG PET Ki parametric maps: application to chemotherapy response in lymphoma

BackgroundTo investigate the use of dynamic radiomics features derived from dual-time-point (DTP-feature) [18F]FDG PET metabolic uptake rate Ki parametric maps to develop a predictive model for response to chemotherapy in lymphoma patients.MethodsWe analyzed 126 lesions from 45 lymphoma patients (responding n = 75 and non-responding n = 51) treated with chemotherapy from two different centers. Static and DTP radiomics features were extracted from baseline static PET images and DTP Ki parametric maps. Spearman’s rank correlations were calculated between static and DTP features to identify features with potential additional information. We first employed univariate analysis to determine correlations between individual features, and subsequently utilized multivariate analysis to derive predictive models utilizing DTP and static radiomics features before and after ComBat harmonization. For multivariate modeling, we utilized both the minimum redundancy maximum relevance feature selection technique and the XGBoost classifier. To evaluate our model, we partitioned the patient datasets into training/validation and testing sets using an 80/20% split. Different metrics for classification including area under the curve (AUC), sensitivity (SEN), specificity (SPE), and accuracy (ACC) were reported in test sets.ResultsVia Spearman’s rank correlations, there was negligible to moderate correlation between 32 out of 65 DTP features and some static features (ρ < 0.7); all the other 33 features showed high correlations (ρ ≥ 0.7). In univariate modeling, no significant difference between AUC of DTP and static features was observed. GLRLM_RLNU from static features demonstrated a strong correlation (AUC = 0.75, p value = 0.0001, q value = 0.0007) with therapy response. The most predictive DTP features were GLCM_Energy, GLCM_Entropy, and Uniformity, each with AUC = 0.73, p value = 0.0001, and q value < 0.0005. In multivariate analysis, the mean ranges of AUCs increased following harmonization. Use of harmonization plus combining DTP and static features was shown to provide significantly improved predictions (AUC = 0.97 ± 0.02, accuracy = 0.89 ± 0.05, sensitivity = 0.92 ± 0.09, and specificity = 0.88 ± 0.05). All models depicted significant performance in terms of AUC, ACC, SEN, and SPE (p < 0.05, Mann–Whitney test).ConclusionsOur results demonstrate significant value in harmonization of radiomics features as well as combining DTP and static radiomics models for predicting response to chemotherapy in lymphoma patients.

Machine learning clinical decision support for interdisciplinary multimodal chronic musculoskeletal pain treatment.

Chronic musculoskeletal pain is a prevalent condition impacting around 20% of people globally; resulting in patients living with pain, fatigue, restricted social and employment capacity, and reduced quality of life. Interdisciplinary multimodal pain treatment programs have been shown to provide positive outcomes by supporting patients modify their behavior and improve pain management through focusing attention on specific patient valued goals rather than fighting pain. Given the complex nature of chronic pain there is no single clinical measure to assess outcomes from multimodal pain programs. Using Centre for Integral Rehabilitation data from 2019-2021 (n = 2,364), we developed a multidimensional machine learning framework of 13 outcome measures across 5 clinically relevant domains including activity/disability, pain, fatigue, coping and quality of life. Machine learning models for each endpoint were separately trained using the most important 30 of 55 demographic and baseline variables based on minimum redundancy maximum relevance feature selection. Five-fold cross validation identified best performing algorithms which were rerun on deidentified source data to verify prognostic accuracy. Individual algorithm performance ranged from 0.49 to 0.65 AUC reflecting characteristic outcome variation across patients, and unbalanced training data with high positive proportions of up to 86% for some measures. As expected, no single outcome provided a reliable indicator, however the complete set of algorithms established a stratified prognostic patient profile. Patient level validation achieved consistent prognostic assessment of outcomes for 75.3% of the study group (n = 1,953). Clinician review of a sample of predicted negative patients (n = 81) independently confirmed algorithm accuracy and suggests the prognostic profile is potentially valuable for patient selection and goal setting. These results indicate that although no single algorithm was individually conclusive, the complete stratified profile consistently identified patient outcomes. Our predictive profile provides promising positive contribution for clinicians and patients to assist with personalized assessment and goal setting, program engagement and improved patient outcomes.

Obstructive sleep apnea detection with nonlinear analysis of speech

Studies on the detection of Obstructive Sleep Apnea (OSA) from speech recording have brought an important innovation to this field. This study is based on the hypothesis that the deterioration of the muscles and tissues in the vocal tract in OSA patients changes the nonlinear properties of voice. In this study, the features that reveal the nonlinear structure of the speech for OSA detection were investigated. The nonlinear characteristics were evaluated for vowels and consonants, and it was tried to find out in which voice group the nonlinear features were more distinctive in OSA. The nonlinear analysis was applied to a wide variety of voice samples having vocal tract components affected by OSA, and OSA/healthy classifications were realized. The results revealed that nonlinear analysis gives considerable findings in OSA detection, and consonants are more successful than vowels. For classifications, K-Nearest Neighbors (KNN) and Support Vector Machine (SVM) classifiers, 5-fold cross validation procedure and Minimum Redundancy Maximum Relevance (MRMR) feature selection method were used. Using the whole dataset, OSA detection performance for vowels was found as 83.5% with KNN, and 96% with SVM for consonants. Additionally, the test process was carried out by using a distinct group, and 82.5% test accuracy was achieved with only 6 features for consonants. The results indicated that the proposed study supports the hypothesis that the nonlinear behavior of vocal tract changes in subjects with OSA, especially for consonants, and has considerable potential for OSA detection as the pre-diagnosis or screening test in clinical use.

Driving drowsiness detection using spectral signatures of EEG-based neurophysiology.

Introduction: Drowsy driving is a significant factor causing dire road crashes and casualties around the world. Detecting it earlier and more effectively can significantly reduce the lethal aftereffects and increase road safety. As physiological conditions originate from the human brain, so neurophysiological signatures in drowsy and alert states may be investigated for this purpose. In this preface, A passive brain-computer interface (pBCI) scheme using multichannel electroencephalography (EEG) brain signals is developed for spatially localized and accurate detection of human drowsiness during driving tasks. Methods: This pBCI modality acquired electrophysiological patterns of 12 healthy subjects from the prefrontal (PFC), frontal (FC), and occipital cortices (OC) of the brain. Neurological states are recorded using six EEG channels spread over the right and left hemispheres in the PFC, FC, and OC of the sleep-deprived subjects during simulated driving tasks. In post-hoc analysis, spectral signatures of the δ, θ, α, and β rhythms are extracted in terms of spectral band powers and their ratios with a temporal correlation over the complete span of the experiment. Minimum redundancy maximum relevance, Chi-square, and ReliefF feature selection methods are used and aggregated with a Z-score based approach for global feature ranking. The extracted drowsiness attributes are classified using decision trees, discriminant analysis, logistic regression, naïve Bayes, support vector machines, k-nearest neighbors, and ensemble classifiers. The binary classification results are reported with confusion matrix-based performance assessment metrics. Results: In inter-classifier comparison, the optimized ensemble model achieved the best results of drowsiness classification with 85.6% accuracy and precision, 89.7% recall, 87.6% F1-score, 80% specificity, 70.3% Matthews correlation coefficient, 70.2% Cohen's kappa score, and 91% area under the receiver operating characteristic curve with 76-ms execution time. In inter-channel comparison, the best results were obtained at the F8 electrode position in the right FC of the brain. The significance of all the results was validated with a p-value of less than 0.05 using statistical hypothesis testing methods. Conclusions: The proposed scheme has achieved better results for driving drowsiness detection with the accomplishment of multiple objectives. The predictor importance approach has reduced the feature extraction cost and computational complexity is minimized with the use of conventional machine learning classifiers resulting in low-cost hardware and software requirements. The channel selection approach has spatially localized the most promising brain region for drowsiness detection with only a single EEG channel (F8) which reduces the physical intrusiveness in normal driving operation. This pBCI scheme has a good potential for practical applications requiring earlier, more accurate, and less disruptive drowsiness detection using the spectral information of EEG biosignals.

SEMG Signal-Based Lower Limb Movements Recognition Using Tunable Q-Factor Wavelet Transform and Kraskov Entropy

BackgroundThe recognition of lower limb movement has a wide range of applications in rehabilitation training, wearable exoskeleton control, and human activity monitoring. Surface electromyography (sEMG) signals can directly reflect the intention of human movement and can be used as the source of lower limb movement recognition. Literature reports have shown that extracting features from sEMG signals is the core of human movement recognition based on sEMG signals. However, how to effectively extract features from the sEMG signal of the lower limbs affected by body gravity is a difficult problem for the recognition of lower limb movement based on the sEMG signal. ObjectivesThe main objective of this paper is to propose an efficient lower limb movement recognition model based on sEMG signals to accurately recognize the four lower limb movements. Methods and resultsWe proposed a novel method of lower limbs movements recognition based on tunable Q-factor wavelet transform (TQWT) and Kraskov entropy (KrEn). Firstly, the sEMG signals of four different lower limb movements from twenty subjects were recorded by seven wearable sEMG signal sensors, and the recorded sEMG signals were denoised by multi-scale principal component analysis (MSPCA). Then, the denoised sEMG signal is decomposed into multiple sub-band signals by TQWT and the KrEn feature is extracted from each sub-band signal. Next, the representative features are selected from the extracted KrEn features by the minimum redundancy maximum relevance (mRMR) feature selection method. Finally, the four lower limb movements are recognized by three machine learning classifiers. Besides, to improve the recognition performance, a majority voting (MV) technology is proposed for the post-processing of decision flow. Experimental results show that the combination of TQWT, KrEn, and MV technology achieved the average recognition accuracy of 98.42% using the linear discriminant analysis (LDA) classifier. ConclusionThe method proposed in this paper can recognize lower limb movements with high accuracy. Compared with existing methods, this method is more advanced and accurate, indicating that it has great application potential in rehabilitation training, wearable exoskeleton control, and daily activity monitoring.

Fluorescence molecular optomic signatures improve identification of tumors in head and neck specimens.

Fluorescence molecular imaging using ABY-029, an epidermal growth factor receptor (EGFR)-targeted, synthetic Affibody peptide labeled with a near-infrared fluorophore, is under investigation for surgical guidance during head and neck squamous cell carcinoma (HNSCC) resection. However, tumor-to-normal tissue contrast is confounded by intrinsic physiological limitations of heterogeneous EGFR expression and non-specific agent uptake. In this preliminary study, radiomic analysis was applied to optical ABY-029 fluorescence image data for HNSCC tissue classification through an approach termed "optomics." Optomics was employed to improve tumor identification by leveraging textural pattern differences in EGFR expression conveyed by fluorescence. The study objective was to compare the performance of conventional fluorescence intensity thresholding and optomics for binary classification of malignant vs. non-malignant HNSCC tissues. Fluorescence image data collected through a Phase 0 clinical trial of ABY-029 involved a total of 20,073 sub-image patches (size of 1.8 × 1.8 mm2) extracted from 24 bread-loafed slices of HNSCC surgical resections originating from 12 patients who were stratified into three dose groups (30, 90, and 171 nanomoles). Each dose group was randomly partitioned on the specimen-level 75%/25% into training/testing sets, then all training and testing sets were aggregated. A total of 1,472 standardized radiomic features were extracted from each patch and evaluated by minimum redundancy maximum relevance feature selection, and 25 top-ranked features were used to train a support vector machine (SVM) classifier. Predictive performance of the SVM classifier was compared to fluorescence intensity thresholding for classifying testing set image patches with histologically confirmed malignancy status. Optomics provided consistent improvement in prediction accuracy and false positive rate (FPR) and similar false negative rate (FNR) on all testing set slices, irrespective of dose, compared to fluorescence intensity thresholding (mean accuracies of 89% vs. 81%, P = 0.0072; mean FPRs of 12% vs. 21%, P = 0.0035; and mean FNRs of 13% vs. 17%, P = 0.35). Optomics outperformed conventional fluorescence intensity thresholding for tumor identification using sub-image patches as the unit of analysis. Optomics mitigate diagnostic uncertainties introduced through physiological variability, imaging agent dose, and inter-specimen biases of fluorescence molecular imaging by probing textural image information. This preliminary study provides a proof-of-concept that applying radiomics to fluorescence molecular imaging data offers a promising image analysis technique for cancer detection in fluorescence-guided surgery.

A Hierarchical Deep Learning-Based Intrusion Detection Architecture for Clustered Internet of Things

The Internet of Things (IoT) system’s ever-expanding attack surface calls for a new intrusion detection system (IDS). These systems may include thousands of wireless devices that need to be protected from cyberattacks. Recent research efforts used machine learning to analyze and identify various attacks and abnormal behavior on IoT systems. Most of these techniques are characterized by low accuracy and they do not scale to today’s IoT-enabled smart cities applications. This article proposes a secure automatic two-levels intrusion detection system (SATIDS) which utilizes the minimum redundancy maximum relevance (MRMR) feature selection technique and an enhanced version of long short-term memory (LSTM) based on an artificial recurrent neural network (RNN) to enhance the IDS performance. SATIDS aims at detecting traffic anomalies with greater accuracy while also reducing the time it takes to perform this task. The proposed algorithm was trained and evaluated using two of the most recent datasets based on realistic data: ToN-IoT and InSDN datasets. The performance analysis of the proposed system proves that it can differentiate between attacks and normal traffic, identify the attack category, and finally define the type of sub-attack with high accuracy. Comparing the performance of the proposed system with the existing IDSs reveals that it outperforms its best rivals from the literature in detecting many types of attacks. It improves accuracy, detection rates, F1-score, and precision. Using 500 hidden and two LSTM layers achieves accuracy of 97.5%, precision of 98.4%, detection rate of 97.9%, and F1-score of 98.05% on ToN-IoT dataset, and precision of 99%, detection rate of 99.6%, and F1-score of 99.3% on InSDN dataset. Finally, SATIDS was applied to an IoT network which utilizes the energy harvesting real-time routing protocol (EHRT). EHRT optimizes the low-energy adaptive clustering hierarchy (LEACH) routing technique using a modified artificial fish swarm algorithm. The integration between the optimized LEACH and the proposed IDS enhances the network lifetime, energy consumption, and security.

OCT-Derived Radiomic Features Predict Anti-VEGF Response and Durability in Neovascular Age-Related Macular Degeneration.

No established biomarkers currently exist for therapeutic efficacy and durability of anti-VEGF therapy in neovascular age-related macular degeneration (nAMD). This study evaluated radiomic-based quantitative OCT biomarkers that may be predictive of anti-VEGF treatment response and durability. Assessment of baseline biomarkers using machine learning (ML) classifiers to predict tolerance to anti-VEGF therapy. Eighty-one participants with treatment-naïve nAMD from the OSPREY study, including 15 super responders (patients who achieved and maintained retinal fluid resolution) and 66 non-super responders (patients who did not achieve or maintain retinal fluid resolution). A total of 962 texture-based radiomic features were extracted from fluid, subretinal hyperreflective material (SHRM), and different retinal tissue compartments of OCT scans. The top 8 features, chosen by the minimum redundancy maximum relevance feature selection method, were evaluated using 4 ML classifiers in a cross-validated approach to distinguish between the 2 patient groups. Longitudinal assessment of changes in different texture-based radiomic descriptors (delta-texture features) between baseline and month 3 also was performed to evaluate their association with treatment response. Additionally, 8 baseline clinical parameters and a combination of baseline OCT, delta-texture features, and the clinical parameters were evaluated in a cross-validated approach in terms of association with therapeutic response. The cross-validated area under the receiver operating characteristic curve (AUC), accuracy, sensitivity, and specificity were calculated to validate the classifier performance. The cross-validated AUC by the quadratic discriminant analysis classifier was 0.75 ± 0.09 using texture-based baseline OCT features. The delta-texture features within different OCT compartments between baseline and month 3 yielded an AUC of 0.78 ± 0.08. The baseline clinical parameters sub-retinal pigment epithelium volume and intraretinal fluid volume yielded an AUC of 0.62 ± 0.07. When all the baseline, delta, and clinical features were combined, a statistically significant improvement in the classifier performance (AUC, 0.81 ± 0.07) was obtained. Radiomic-based quantitative assessment of OCT images was shown to distinguish between super responders and non-super responders to anti-VEGF therapy in nAMD. The baseline fluid and SHRM delta-texture features were found to be most discriminating across groups.

Improving pre-movement pattern detection with filter bank selection

Objective. Pre-movement decoding plays an important role in detecting the onsets of actions using low-frequency electroencephalography (EEG) signals before the movement of an upper limb. In this work, a binary classification method is proposed between two different states. Approach. The proposed method, referred to as filter bank standard task-related component analysis (FBTRCA), is to incorporate filter bank selection into the standard task-related component analysis (STRCA) method. In FBTRCA, the EEG signals are first divided into multiple sub-bands which start at specific fixed frequencies and end frequencies that follow in an arithmetic sequence. The STRCA method is then applied to the EEG signals in these bands to extract CCPs. The minimum redundancy maximum relevance feature selection method is used to select essential features from these correlation patterns in all sub-bands. Finally, the selected features are classified using the binary support vector machine classifier. A convolutional neural network (CNN) is an alternative approach to select canonical correlation patterns. Main Results. Three methods were evaluated using EEG signals in the time window from 2 s before the movement onset to 1 s after the movement onset. In the binary classification between a movement state and the resting state, the FBTRCA achieved an average accuracy of 0.8968 ± 0.0847 while the accuracies of STRCA and CNN were 0.8228 ± 0.1149 and 0.8828 ± 0.0917, respectively. In the binary classification between two actions, the accuracies of STRCA, CNN, and FBTRCA were 0.6611 ± 0.1432, 0.6993 ± 0.1271, 0.7178 ± 0.1274, respectively. Feature selection using filter banks, as in FBTRCA, produces comparable results to STRCA. Significance. The proposed method provides a way to select filter banks in pre-movement decoding, and thus it improves the classification performance. The improved pre-movement decoding of single upper limb movements is expected to provide people with severe motor disabilities with a more natural, non-invasive control of their external devices.