Filter-based Method Research Articles

An Efficient Filter and Wrapper based Selection Methods along With Random Forest and Support Vector Machines Classification Technique in Health Care System

Health care Management System (HMS) is a key to successful management of any health care industry. Health care management systems have so many research dimensions such as identifying disease and diagnostic, drug discovery manufacturing, Bioinformatics’ problem, personalized treatments, Patient image analysis and so on. Heart Disease Prediction (HDP) is a process of identifying heart disease in advance and recognizes patient health condition by applying techniques on patient heart related symptoms. Now a day’s the problem of identifying heart diseases is solved by machine learning techniques. In this paper we construct a heart disease prediction method using combined feature selection and classification machine learning techniques. According to the existing study the one of the main difficult in heart disease prediction system is that the available data in open sources are not properly recorded the necessary characteristics and there is some lagging in finding the useful features from the available features. The process of removing inappropriate features from an available feature set while preserving sufficient classification accuracy is known as feature selection. A methodology is proposed in this paper that consists of two phases: Phase one employs two broad categories of feature selection techniques to identify the efficient feature sets and it is given to the input of our second phase such as classification. In this work we will concentrate on filter-based method for feature selection such as Chi-square, Fast Correlation Based Filter (FCBF), Gini Index (GI), RelifeF, and wrapper-based method for feature selection such as Backward Feature Elimination (BFE), Exhaustive Feature Selection (EFS), Forward Feature Selection (FFS), and Recursive Feature Elimination (RFE). The UCI heart disease data set is used to evaluate the output in this study. Finally, the proposed system's performance is validated by various experiments setups.

Color filter-based gait silhouette extraction method in dynamic background

Color filter-based gait silhouette extraction method in dynamic background

A multi-objective mutation-based dynamic Harris Hawks optimization for botnet detection in IoT

The increasing trend toward using the Internet of Things (IoT) increased the number of intrusions and intruders annually. Hence, the integration, confidentiality, and access to digital resources would be threatened continually. The significance of security implementation in digital platforms and the need to design defensive systems to discover different intrusions made the researchers study updated and effective methods, such as Botnet Detection for IoT systems. Many problem space features and network behavior unpredictability made the Intrusion Detection System (IDS) the main problem in maintaining computer networks' security. Furthermore, many insignificant features have turned the feature selection (FS) problem into a vast IDS aspect. This paper introduces a novel binary multi-objective dynamic Harris Hawks Optimization (HHO) enhanced with mutation operator (MODHHO) and applies it to Botnet Detection in IoT. Afterward, the Feature Selection (FS) is undertaken, and the K-Nearest Neighbor (KNN), Support Vector Machine (SVM), Multilayer Perceptron (MLP), and Decision Tree (DT) classifiers are used to estimate the potential of the selected features in the precise detection of intrusions. The simulation results illustrated that the MODHHO algorithm performs well in Botnet Detection in IoT and is preferred to other approaches in its performance metrics. Besides, the computational complexity analysis results suggest that the MODHHO algorithm's overhead is more optimal than similar approaches. The MODHHO algorithm has performed better in comparison with other compared algorithms in all 5 data sets. In contrast with the machine learning methods of the proposed model in all five data sets, it has had a better error rate according to the AUC, G-mean, and TPR criteria. And according to the comparison made with filter-based methods, it has performed almost better in three datasets.

Multiparticle Kalman filter for object localization in symmetric environments

This study considers the object localization problem and proposes a novel multiparticle Kalman filter to solve it in complex and symmetric environments. Two well-known classes of filtering algorithms to solve the localization problem are Kalman filter-based methods and particle filter-based methods. We consider these classes, demonstrate their complementary properties, and propose a novel filtering algorithm that takes the best from two classes. We evaluate the multiparticle Kalman filter in symmetric and noisy environments. Such environments are especially challenging for both classes of classical methods. We compare the proposed approach with the particle filter since only this method is feasible if the initial state is unknown. In the considered challenging environments, our method outperforms the particle filter in terms of both localization error and runtime.

Intelligent Mesh Cluster Algorithm for Device-Free Localization in Wireless Sensor Networks

Device-free localization (DFL) is a technology designed to determine the positions of targets without the need for them to carry electronic devices. It achieves this by analyzing the shadowing effects of radio links within wireless sensor networks (WSNs). However, obtaining high precision in DFL often results in increased energy consumption, severe electromagnetic interference, and other challenges that impact positioning accuracy. Most DFL schemes for accurate tracking require substantial memory and computing resources, which make them unsuitable for resource-constrained applications. To address these challenges, we propose an intelligent mesh cluster (IMC) algorithm that achieves accurate tracking by adaptively activating a subset of wireless links. This approach not only reduces electromagnetic interference but also saves energy. The IMC algorithm leverages geometric objects, such as meshes and mesh clusters formed by wireless links, to achieve low computational complexity. By scanning a subset of mesh cluster-related wireless links near the DFL target, the algorithm significantly reduces the computational requirements. The target’s location estimate is determined based on the connection information among the mesh clusters. We conducted numerous simulations to evaluate the performance of the IMC algorithm. The results demonstrate that the IMC algorithm outperforms grid-based and particle filter-based DFL methods, confirming its effectiveness in achieving accurate and efficient localization.

Development of Blueberry-Derived Extracellular Nanovesicles for Immunomodulatory Therapy.

Over the past decade, there has been a significant expansion in the development of plant-derived extracellular nanovesicles (EVs) as an effective drug delivery system for precision therapy. However, the lack of effective methods for the isolation and characterization of plant EVs hampers progress in the field. To solve a challenge related to systemic separation and characterization in the plant-derived EV field, herein, we report the development of a simple 3D inner filter-based method that allows the extraction of apoplastic fluid (AF) from blueberry, facilitating EV isolation as well as effective downstream applications. Class I chitinase (PR-3) was found in blueberry-derived EVs (BENVs). As Class I chitinase is expressed in a wide range of plants, it could serve as a universal marker for plant-derived EVs. Significantly, the BENVs exhibit not only higher drug loading capacity than that reported for other EVs but also possess the ability to modulate the release of the proinflammatory cytokine IL-8 and total glutathione in response to oxidative stress. Therefore, the BENV is a promising edible multifunctional nano-bio-platform for future immunomodulatory therapies.

Filter-based feature selection methods in the presence of missing data for medical prediction models

Filter-based feature selection methods in the presence of missing data for medical prediction models

Drift-Free Joint Angle Calculation Using Inertial Measurement Units without Magnetometers: An Exploration of Sensor Fusion Methods for the Elbow and Wrist.

Joint angles of the lower extremities have been calculated using gyroscope and accelerometer measurements from inertial measurement units (IMUs) without sensor drift by leveraging kinematic constraints. However, it is unknown whether these methods are generalizable to the upper extremity due to differences in motion dynamics. Furthermore, the extent that post-processed sensor fusion algorithms can improve measurement accuracy relative to more commonly used Kalman filter-based methods remains unknown. This study calculated the elbow and wrist joint angles of 13 participants performing a simple ≥30 min material transfer task at three rates (slow, medium, fast) using IMUs and kinematic constraints. The best-performing sensor fusion algorithm produced total root mean square errors (i.e., encompassing all three motion planes) of 6.6°, 3.6°, and 2.0° for the slow, medium, and fast transfer rates for the elbow and 2.2°, 1.7°, and 1.5° for the wrist, respectively.

Tracking-Aided Respiration Detection Using Radar During Large-Scale Body Movements

Recently, human respiration detection using microwave radars has attracted significant interest as a noninvasive and noncontact form of measurement. Most prior research has focused on respiration detection of the subjects who are stationary or in random body movement. So far, few reported studies have been carried out on detecting the subjects in large-scale body movement. In this paper, we present a tracking-aided method, which integrates the Kalman filter-based tracking method and a range-spread human subject model, to detect the subjects moving across multiple range bins. The experimental results validate the correctness of the range-spread model and the effectiveness of the proposed method.

SDIF-CNN: Stacking deep image features using fine-tuned convolution neural network models for real-world malware detection and classification

The detection of malware is a complex problem in the area of Internet security. Developing a malware defense system that is less costly to detect large-scale malware is needed. This paper proposes a novel malware detection and classification architecture based on image visualization as SDIF-CNN: Stacking deep image features using fine-tuned convolution neural networks. The hybrid methodology of transfer learning as fine-tuning and feature extractor of deep convolution neural network models is designed. At first, the pre-trained VGG16 CNN model is deeply fine-tuned with different hyperparameters, including the number of layers, learning rate, momentum, etc. The transfer learning-based fine-tuned VGG16 model is used as a feature extractor along with the three similar pre-trained CNN models, VGG19, ResNet50, and InceptionV3, to obtain the diverse feature map. The extracted features are horizontally concatenated to construct a single feature map. The different feature selection methodologies, including filter-based methods and embedded methods, such as linear regression and random forest, are designed to discard the irrelevant features from a stacked feature map. After that, this study uses six machine learning and deep learning classifiers- K-Nearest Neighbor (K-NN), Support Vector Machine (SVM), Random Forest (RF), Multi-Layer Perceptron (MLP), Extra Tree (ET), and Gaussian Naive Bayes (GNB) by using the stacked feature map as a training feature vector. The hyperparameter optimization of the MLP model as the best classifier is performed using a randomized search algorithm to devise an optimal classifier. The experiments are performed using a publicly benchmarked MalImg dataset of 9339 images from 25 families. The model is also validated on real-world and packed malicious programs to prove the generalization of the proposed methodology in detecting real-world malware. In the proposed system, the MLP model obtained the best performance results as 98.55% accuracy, 99% precision, 99% recall, and 99% F1-score for MalImg datasets, and accuracy of 94.78% for real-world malware datasets. The proposed methodology is resilient to commonly used obfuscation techniques and does not depend upon code disassembly, reverse engineering analysis, and highly resource-intensive dynamic analysis.

A Multifactor Eigenvector Spatial Filtering-Based Method for Resolution-Enhanced Snow Water Equivalent Estimation in the Western United States

Accurate snow water equivalent (SWE) products are vital for monitoring hydrological processes and managing water resources effectively. However, the coarse spatial resolution (typically at 25 km from passive microwave remote sensing images) of the existing SWE products cannot meet the needs of explicit hydrological modeling. Linear regression ignores the spatial autocorrelation (SA) in the variables, and the measure of SA in the data assimilation algorithm is not explicit. This study develops a Resolution-enhanced Multifactor Eigenvector Spatial Filtering (RM-ESF) method to estimate daily SWE in the western United States based on a 6.25 km enhanced-resolution passive microwave record. The RM-ESF method is based on a brightness temperature gradience algorithm, incorporating not only factors including geolocation, environmental, topographical, and snow features but also eigenvectors generated from a spatial weights matrix to take SA into account. The results indicate that the SWE estimation from the RM-ESF method obviously outperforms other SWE products given its overall highest correlation coefficient (0.72) and lowest RMSE (56.70 mm) and MAE (43.88 mm), compared with the AMSR2 (0.33, 131.38 mm, and 115.45 mm), GlobSnow3 (0.50, 100.03 mm, and 83.58 mm), NCA-LDAS (0.48, 98.80 mm, and 81.94 mm), and ERA5 (0.65, 67.33 mm, and 51.82 mm), respectively. The RM-ESF model considers SA effectively and estimates SWE at a resolution of 6.25 km, which provides a feasible and efficient approach for SWE estimation with higher precision and finer spatial resolution.

LOKALISASI PLAT NOMOR DENGAN JAVA MENGGUNAKAN DETEKSI TEPI CANNY DAN FILTER MORFOLOGI

This study aims to introduce an automatic vehicle identification system based on vehicle licenses and license plate recognition. This Automatic Number Plate Recognition (ANPR) system software uses a series of image processing algorithms captured by a camera with a high degree of accuracy and has the efficiency to recognize license plates and then identify vehicles from datasets stored on a Personal Computer (PC). This study uses an algorithm for license plate localization by implementing the Canny edge detection algorithm and a morphological filter.The methodology used is to develop a Number Plate (NP) localization method including the following steps: (1) Color conversion, image enhancement (using sharpness filters and gamma correction), Canny edge detection and alignment (skeletonization); stage (2) morphological filter-based method to determine the Region of Interest (ROI), namely a rectangular box containing a Number Plate (NP). A rectangular bounding box is drawn across the area in the original image that contains the NP after the NP are localized. The bounding box is equal to the width of the image plus one pixel and the height of the NP plus one pixel.The results showed that this method succeeded in achieving a localization rate of about 72.0%. The method used for license plate localization is able to localize license plates with clear boundaries and unambiguous separation of other related components. This research can be further developed to include more functions so that it can be used by all elements of society from researchers to law enforcers.

Subspace-based modal identification and uncertainty quantification from video image flows

Vibration measurements for structural health monitoring (SHM) by operational modal analysis (OMA) are classically obtained from sensors that are embedded in or physically attached to the monitored structure, like accelerometers or strain gauges. However, the setup time of these sensors and their restricted number and space coverage limit their monitoring capabilities. Video image-based sensing methods can overcome these shortcomings. With adequate image processing methods, motion signals are extracted from video image flows, which are then processed by system identification methods to estimate modal parameters. In this way, the pixels in selected regions of interest within the images act as a dense network of contactless sensors distributed over the whole structure. In this paper, the efficiency of this video-based approach is demonstrated with laboratory experiments on a cantilever beam, in particular, by evaluating its capability for detecting weak damages mimicked by slight mass modifications. To this end, the steerable filter-based method (ST), that recovers displacements from local phase, is first extended to overcome its motion limitation of one pixel size. Then, the performance of the improved motion extraction method is compared with two other well-established methods in the context of OMA, where natural frequencies, damping ratios and mode shapes with high spatial resolution are estimated together with their uncertainty bounds using covariance-driven subspace identification. The compared methods are evaluated with the help of reference laser displacement measurements as well as a finite element model of the beam, revealing differences in the accuracy of the estimated mode shapes depending on the chosen method for motion extraction. Finally, aiming to investigate early structural damage detection, experiments are carried out under small structural changes and the results are compared to a reference state with the help of estimated uncertainties. Small but statistically significant changes in the modal parameters are detected, showing the potential of the vision based framework for SHM.

An investigation of rain streak removal models based on expert experience and deep learning

Computer vision technology has a wide range of applications in today's society, and image rain removal is of great importance in outdoor vision capture. Today's image de-rain techniques are divided into video de-rain, and image de-rain, with the image de-rain task being more difficult than the video de-rain task due to the lack of a time factor. Current image rain removal methods are divided into three main types: filter-based methods, a priori knowledge-based methods and deep learning methods. Although these methods can achieve the image rain removal requirements to a certain extent, there is still no highly generalized method that can better solve the image rain removal problem in all cases. This paper first considers a filter-based approach, which takes less time to run but is difficult to remove cleanly for complex rain streaks. Secondly, this paper examines an a priori knowledge-based approach, which requires the study of how rain images are constructed as a priori knowledge and then uses the existing prior knowledge to remove rain from the images. This approach has a high reliance on prior knowledge and poor generalization. Finally, this paper investigates a deep learning-based method that requires a large number of supervised samples for training and has a better rain removal effect, but ignores the prior knowledge of rain streaks and is prone to overfitting. Based on these three methods we collected some data and experimental results in this field and summarized and analyzed them, giving reasons for the strengths and weaknesses of these three models and presenting new perspectives for future improvements in image de-rain methods.

Disturbance observer based actor-critic learning control for uncertain nonlinear systems

This paper investigates the disturbance observer based actor-critic learning control for a class of uncertain nonlinear systems in the presence of unmodeled dynamics and time-varying disturbances. The proposed control algorithm integrates a filter-based design method with actor-critic learning architecture and disturbance observer to circumvent the unmodeled dynamic and the time-varying disturbance. To be specific, the actor network is employed to estimate the unknown system dynamic, the critic network is developed to evaluate the control performance, and the disturbance observer is leveraged to provide efficient estimation of the compounded disturbance which includes the time-varying disturbance and the actor-critic network approximation error. Consequently, high-gain feedback is avoided and the improved tracking performance can be expected. Moreover, a composite weight adaptation law for actor network is constructed by utilizing two types of signals, the cost function and the modeling error. Eventually, theoretical analysis demonstrates that the developed controller can guarantee bounded stability. Extensive simulations and experiments on a robot manipulator are implemented to validate the performance of the resulted control strategy.

Genomic characterisation of a blaKPC-2- and mcr-10-co-harbouring Enterobacter kobei isolate with high-level resistance to colistin and carbapenems

The emergence and spread of colistin resistance in carbapenem-resistant Enterobacteriaceae pose a serious threat to human and animal health. This work aimed to characterize the genetic features of antimicrobial resistance of a carbapenem and colistin resistant Enterobacter kobei SCLZS19 from hospital sewage by using whole genome sequencing. Antimicrobial susceptibility tests were performed by using the disk diffusion method. Whole genome sequencing of SCLZS19 was carried out on the HiSeq 2000 combined with PacBio RSII platforms. Sequence type, plasmid incompatibility types, resistance genes, and insertion elements were identified using the MLST, PlasmidFinder, ResFinder, and ISfinder, respectively. Conjugation assays were performed using both broth- and filter-based methods with the azide-resistant Escherichia coli J53 as the recipient. The function of the mcr-9-like variant was determined by gene cloning. E. kobei SCLZS19 has a 4,862,177-bp circular chromosome, and nine circular plasmids ranging in size from 4,120 bp to 282,472 bp. It carries 11 antibiotic resistance genes (ARGs), and 10 of them are located on plasmids. The colistin resistance gene mcr-10 is located on a 118,766-bp non-transferable IncF (Y3:A-:B-) plasmid. The carbapenemase gene blaKPC-2 is carried by a self-transmissible IncP6 plasmid which is epidemic in China. In addition, SCLZS19 also carries a mcr-9-like variant on a IncHI2 (ST1) plasmid. Cloning assay shows that the mcr-9-like variant does not mediate colistin resistance in E. coli DH5α. The findings highlight that carbapenem- and colistin-resistant Enterobacterales from water environment may serve as a reservoir for clinically significant ARGs, and continuous surveillance is required.

Towards Robust Visual Tracking for Unmanned Aerial Vehicle with Spatial Attention Aberration Repressed Correlation Filters

In recent years, correlation filtering has been widely used in the field of UAV target tracking for its high efficiency and good robustness, even on a common CPU. However, the existing correlation filter-based tracking methods still have major problems when dealing with challenges such as fast moving targets, camera shake, and partial occlusion in UAV scenarios. Furthermore, the lack of reasonable attention mechanism for distortion information as well as background information prevents the limited computational resources from being used for the part of the object most severely affected by interference. In this paper, we propose the spatial attention aberration repressed correlation filter, which models the aberrations, makes full use of the spatial information of aberrations and assigns different attentions to them, and can better cope with these challenges. In addition, we propose a mechanism for the intermittent learning of the global context to balance the efficient use of limited computational resources and cope with various complex scenarios. We also tested the mechanism on challenging UAV benchmarks such as UAVDT and Visdrone2018, and the experiments show that SAARCF has better performance than state-of-the-art trackers.

Feasibility of local interpretable model-agnostic explanations (LIME) algorithm as an effective and interpretable feature selection method: comparative fNIRS study.

Many feature selection methods have been evaluated in functional near-infrared spectroscopy (fNIRS)-related studies. The local interpretable model-agnostic explanation (LIME) algorithm is a feature selection method for fNIRS datasets that has not yet been validated; the demand for its validation is increasing. To this end, we assessed the feature selection performance of LIME for fNIRS datasets in terms of classification accuracy. A comparative analysis was conducted for the benchmark (classification accuracy obtained without applying any feature selection method), LIME, two filter-based methods (minimum-redundancy maximum-relevance and t-test), and one wrapper-based method (sequential forward selection). To ensure the fairness and reliability of the performance evaluation, several open-access fNIRS datasets were used. The analysis revealed that LIME greatly outperformed the other feature selection methods in most cases and could achieve a statistically significantly better classification accuracy than that of the benchmark methods. These findings implied the effectiveness of LIME as a feature selection approach for fNIRS datasets.

A Multi-Scale Gabor Filter-Based Method for Enhancing Video Images in Distance Education

A Multi-Scale Gabor Filter-Based Method for Enhancing Video Images in Distance Education

Detection of depression and anxiety in the perinatal period using Marine Predators Algorithm and kNN

Undiagnosed prenatal anxiety and depression have the potential to worsen and have an adverse effect on both the mother and the infant. Although the diagnosis is made by specialist doctors, it is unclear which parameters are more effective. Especially in medicine, it is crucial to diagnose disease with high accuracy. For this reason, in this study, a questionnaire study was first conducted on pregnant women, and real original data were collected. Then, the Marine Predators Algorithm (MPA), one of the current metaheuristic algorithms inspired by nature, was combined with K-Nearest Neighbors (kNN) to determine high-priority features in the collected data. As a result, five of the 147 features selected by the proposed method were determined as high priority and approved by the doctors. In addition, the proposed method is compared with the Chi-square method, which is one of the filter-based feature selection methods. Thanks to the proposed feature selection method based on MPA and kNN, it has been observed that the classification gives more successful results in a shorter time with 98.11% success, and the model supports the diagnosis stage of the doctors.