Terms Of Several Performance Metrics Research Articles

Predicting pedestrian-involved crash severity using inception-v3 deep learning model

This research leverages a novel deep learning model, Inception-v3, to predict pedestrian crash severity using data collected over five years (2016–2021) from Louisiana. The final dataset incorporates forty different variables related to pedestrian attributes, environmental conditions, and vehicular specifics. Crash severity was classified into three categories: fatal, injury, and no injury. The Boruta algorithm was applied to determine the importance of variables and investigate contributing factors to pedestrian crash severity, revealing several associated aspects, including pedestrian gender, pedestrian and driver impairment, posted speed limits, alcohol involvement, pedestrian age, visibility obstruction, roadway lighting conditions, and both pedestrian and driver conditions, including distraction and inattentiveness. To address data imbalance, the study employed Random Under Sampling (RUS) and the Synthetic Minority Oversampling Technique (SMOTE). The DeepInsight technique transformed numeric data into images. Subsequently, five crash severity prediction models were developed with Inception-v3, considering various scenarios, including original, under-sampled, over-sampled, a combination of under and over-sampled data, and the top twenty-five important variables. Results indicated that the model applying both over and under sampling outperforms models based on other data balancing techniques in terms of several performance metrics, including accuracy, sensitivity, precision, specificity, false negative ratio (FNR), false positive ratio (FPR), and F1-score. This model achieved prediction accuracies of 93.5%, 77.5%, and 85.9% for fatal, injury, and no injury categories, respectively. Additionally, comparative analysis based on several performance metrics and McNemar’s tests demonstrated that the predictive performance of the Inception-v3 deep learning model is statistically superior compared to traditional machine learning and statistical models. The insights from this research can be effectively harnessed by safety professionals, emergency service providers, traffic management centers, and vehicle manufacturers to enhance their safety measures and applications.

Multi-Criteria Decision-Making for Energy Management in Smart Homes Using Hybridized Neuro-Fuzzy Approach

The necessity for smart energy oversight solutions has arisen in response to the rising popularity of energy-efficient home automation and other energy-saving technologies. Optimizing smart home energy use using multi-criteria decision-making (MCDM) is a proven methodology. However, the procedure for making decisions and MCDM’s capacity to handle various criteria are typically limiting factors. Hybrid methods, which integrate multiple decision-making approaches like Fuzzy Logic (FL) and Modular Neural Networks (MNN), could potentially be able to circumvent these restrictions and boost energy management systems’ efficacy and precision. This investigation presents a hybrid Neuro-Fuzzy (H-NF) method for MCDM in regulating energy for smart homes by combining FL with an MNN. The suggested approach would optimize energy use in smart homes by considering several parameters, notably cost, ease of use, and environmental effects. In addition, this study aims to examine how the H-NF model fares in comparison to other methods of making important decisions in terms of several performance metrics. The suggested hybridized approach has the potential to deliver more precise and effective decision-making processes for energy management in smart homes, allowing users to optimize their energy consumption while preserving comfort and lowering environmental impact.

Optimizing jointly mining decision and resource allocation in a MEC-enabled blockchain networks

In this paper, several recently published metaheuristic algorithms are adapted to optimize the NP-hard problem of jointly mining decision and resource allocation inmobile edge computing (MEC) enabled blockchain networks under two different encoding schemes. The first scheme represents individuals in a way that incorporates the mining decisions, transmission power, and computing resources of all miners for each individual, with mining decisions determined by a binary vector whose values indicate whether miners partake in mining or not. While, the second scheme makes each individual accountable for the transmission power and computing resources of each participant miner, treating all individuals as a singular solution to the problem. Then the Nutcracker optimization algorithm and gradient-based optimizer are modified to propose two robust variants, MNOA and MGBO, respectively. We then combine MNOA and MGBO to create HNOA, which further optimizes the mining decision and resource allocation in this problem. HNOA and other variants are validated using nine instances with a range of 150 to 600 miners. HNOA is also compared to several competing optimizers to demonstrate its efficacy in terms of several performance metrics. The experimental findings show the superiority of the proposed algorithm.

An Efficient Evolution-Based Technique for Moving Target Search with Unmanned Aircraft Vehicle: Analysis and Validation

Recent advances in technology have led to a surge in interest in unmanned aerial vehicles (UAVs), which are remote-controlled aircraft that rely on cameras or sensors to gather information about their surroundings during flight. A UAV requires a path-planning technique that can swiftly recalculate a viable and quasi-optimal path in flight if a new obstacle or hazard is recognized or if the target is moved during the mission. In brief, the planning of UAV routes might optimize a specific problem determined by the application, such as the moving target problem (MTP), flight time and threats, or multiobjective navigation. The complexity of MTP ranges from NP-hard to NEXP-complete because there are so many probabilistic variables involved. Therefore, it is hard to detect a high-quality solution for this problem using traditional techniques such as differential calculus. Therefore, this paper hybridizes differential evolution (DE) with two newly proposed updating schemes to present a new evolution-based technique named hybrid differential evolution (HDE) for accurately tackling the MTP in a reasonable amount of time. Using Bayesian theory, the MTP can be transformed into an optimization problem by employing the target detection probability as the fitness function. The proposed HDE encodes the search trajectory as a sequence of UAV motion pathways that evolve with increasing the current iteration for finding the near-optimal solution, which could maximize this fitness function. The HDE is extensively compared to the classical DE and several rival optimizers in terms of several performance metrics across four different scenarios with varying degrees of difficulty. This comparison demonstrates the proposal’s superiority in terms of the majority of used performance metrics.

Xception-Fractalnet: Hybrid Deep Learning Based Multi-Class Classification of Alzheimer’s Disease

Neurological disorders such as Alzheimer’s disease (AD) are very challenging to treat due to their sensitivity, technical challenges during surgery, and high expenses. The complexity of the brain structures makes it difficult to distinguish between the various brain tissues and categorize AD using conventional classification methods. Furthermore, conventional approaches take a lot of time and might not always be precise. Hence, a suitable classification framework with brain imaging may produce more accurate findings for early diagnosis of AD. Therefore in this paper, an effective hybrid Xception and Fractalnet-based deep learning framework are implemented to classify the stages of AD into five classes. Initially, a network based on Unet++ is built to segment the tissues of the brain. Then, using the segmented tissue components as input, the Xception-based deep learning technique is employed to extract high-level features. Finally, the optimized Fractalnet framework is used to categorize the disease condition using the acquired characteristics. The proposed strategy is tested on the Alzheimer’s Disease Neuroimaging Initiative (ADNI) dataset that accurately segments brain tissues with a 98.45% of dice similarity coefficient (DSC). Additionally, for the multiclass classification of AD, the suggested technique obtains an accuracy of 99.06%. Moreover, ANOVA statistical analysis is also used to evaluate if the groups are significant or not. The findings show that the suggested model outperforms various state-of-the-art methods in terms of several performance metrics.

A genetic algorithm inspired optimized cluster head selection method in wireless sensor networks

• A CH selection optimized based on GA (NCOGA) for the heterogeneous WSN is proposed. • A comprehensive discussion on related work is presented which highlights the research gaps. • The proposed algorithm is discussed in a step-by-step manner. • Rigorous simulations and analysis are presented. • Results are validated against the most recent state-of-the-art algorithms. In this paper, an optimized cluster head (CH) selection method based on genetic algorithm (NCOGA) is proposed which uses the adaptive crossover and binary tournament selection methods to prolong the lifetime of a heterogeneous wireless sensor network (WSN). The novelty of the proposed algorithms is the integration of multiple parameters for the CH selection in a heterogeneous WSN. NCOGA formulates fitness parameters by integrating multiple parameters like the residual energy, initial energy, distance to the sink, number of neighbors surrounded by a node, load balancing factor, and communicating mode decider (CMD). The parameters for load balancing and CMD are utilized to discover out the best candidate to be selected as a relay CH and for deciding the mode of communication (single or multi-hop) of CH. Further, these parameters are useful in avoiding hot-spot problem in the network. The working of the NCOGA starts based on the criteria “consider only those nodes which have energy higher than the pre-defined threshold energy”. This criterion of nodes selection makes the NCOGA more efficient and quickly convergent. Extensive computer simulations are conducted to determine the effectiveness of the NCOGA. Simulation results reveal that the proposed NCOGA outperforms the state-of-the-art optimization algorithms based on GA in terms of several performance metrics, specifically, stability period, residual energy, network lifetime, and throughput.

Influenza-like Illness Detection from Arabic Facebook Posts Based on Sentiment Analysis and 1D Convolutional Neural Network

The recent large outbreak of infectious diseases, such as influenza-like illnesses and COVID-19, has resulted in a flood of health-related posts on the Internet in general and on social media in particular, in a wide range of languages and dialects around the world. The obvious relationship between the number of infectious disease cases and the number of social media posts prompted us to consider how we can leverage such health-related content to detect the emergence of diseases, particularly influenza-like illnesses, and foster disease surveillance systems. We used Algerian Arabic posts as a case study in our research. From data collection to content classification, a complete workflow was implemented. The main contributions of this work are the creation of a large corpus of Arabic Facebook posts based on Algerian dialect and the proposal of a new classification model based on sentiment analysis and one-dimensional convolutional neural networks. The proposed model categorizes Facebook posts based on the users’ feelings. To counteract data imbalance, two techniques have been considered, namely, SMOTE and random oversampling (ROS). Using a 5-fold cross-validation, the proposed model outperformed other baseline and state-of-the-art models such as SVM, LSTM, GRU, and BiLTSM in terms of several performance metrics.

A Review on Major Image and Video Steganography Techniques

With the growing expansion of the intensive sharing of multimedia content and secret communications, data concealing techniques have become increasingly crucial. Steganography is a mechanism for sharing data covertly and securely. It is the science of embedding hidden information into cover media by altering the cover image in a way that is difficult to detect by human eyes. Audio, video and image files can all benefit from steganography techniques. Video steganography is the process of hiding secret information in a video file. Video Steganography is concealing a secret message, which might be a secret text message or an image within a bigger one, in such a way that an unwelcome person cannot determine the presence of the concealed message just by looking at it. There are several Steganography strategies for hiding hidden information in videos, which are further detailed in this paper, along with some of the research efforts done in some disciplines under video steganography by some authors.This article discusses the advancements in the subject of image and video steganography, as well as a comparison of its many applications and methodologies. It is critical to secure digital information while communicating over the internet in today’s digitally-driven society. One of the techniques utilised for this is steganography. The importance of steganography and the numerous types of steganography are discussed in this study. It also clarified the vocabulary for the general model of digital steganography. The traditional steganography approach of most minor significant bit substitution is explored. Furthermore, in terms of several performance metrics, a comparison is made between the conventional technique and other techniques for covering media quality and imperceptibility. The size of the hidden message that can be embedded in the image is estimated.

An experimental evaluation of localization methods used in wireless sensor networks

<span class="fontstyle0">The problem of localization in wireless sensor networks has received considerable attention from researchers over the past decades. Several methods and algorithms have been proposed to solve this problem. The effectiveness of these algorithms depends on the accuracy of the estimated positions and the information required to calculate the coordinates. In this paper, we propose to evaluate four of the most commonly used localization methods in sensor networks. Our study considers a mathematical description of the studied methods in order to evaluate their complexity, and then a practical implementation on the simulation tool Cooja. We evaluate the performance of the studied methods as a function of the number of deployed sensor nodes and their degree of mobility in terms of several performance metrics. The objective is to reveal the most suitable localization method for a particular case of deployment. Improvement proposals are also provided to improve the most relevant localization method for the investigated study.</span> <br /><br />

A robust framework for shoulder implant X-ray image classification

PurposeMedical images are increasingly popular; therefore, the analysis of these images based on deep learning helps diagnose diseases become more and more essential and necessary. Recently, the shoulder implant X-ray image classification (SIXIC) dataset that includes X-ray images of implanted shoulder prostheses produced by four manufacturers was released. The implant's model detection helps to select the correct equipment and procedures in the upcoming surgery.Design/methodology/approachThis study proposes a robust model named X-Net to improve the predictability for shoulder implants X-ray image classification in the SIXIC dataset. The X-Net model utilizes the Squeeze and Excitation (SE) block integrated into Residual Network (ResNet) module. The SE module aims to weigh each feature map extracted from ResNet, which aids in improving the performance. The feature extraction process of X-Net model is performed by both modules: ResNet and SE modules. The final feature is obtained by incorporating the extracted features from the above steps, which brings more important characteristics of X-ray images in the input dataset. Next, X-Net uses this fine-grained feature to classify the input images into four classes (Cofield, Depuy, Zimmer and Tornier) in the SIXIC dataset.FindingsExperiments are conducted to show the proposed approach's effectiveness compared with other state-of-the-art methods for SIXIC. The experimental results indicate that the approach outperforms the various experimental methods in terms of several performance metrics. In addition, the proposed approach provides the new state of the art results in all performance metrics, such as accuracy, precision, recall, F1-score and area under the curve (AUC), for the experimental dataset.Originality/valueThe proposed method with high predictive performance can be used to assist in the treatment of injured shoulder joints.

All-in-one: Toward hybrid data collection and energy saving mechanism in sensing-based IoT applications

Big data collection and storage have become one of the most obvious challenge in this era. Indeed, much of that data is collected thanky to a huge number of connected devices in sensing-based IoT applications. Thus, in order to deal with data growth in such applications, researchers have focused on data reduction approach as an efficient solution for minimizing the amount of data collection and saving the limited sensor energy in such networks. Mainly, data reduction approach relies on various kinds of data processing techniques such that aggregation, compression, prediction, clustering, sensing frequency adaptation and spatial-temporal correlation. However, each of those techniques has its own advantages and disadvantages regarding sensor energy saving, data reduction ratio, data accuracy, complexity, etc. In this paper, we propose a hybrid data collection and energy saving mechanism, called All-in-One, for sensing-based IoT applications. The proposed mechanism takes advantages from existing data reduction techniques while optimizing various performance metrics. All-in-One relies on the cluster network architecture and works on three main phases: on-period, in-period and in-node. The first phase, e.g. on-period, allows each sensor node to search the similarity among its periodic collected data then to reduce its data transmission to the Cluster-Head (CH) by applying either data aggregation, compression or prediction technique. The second phase, e.g. in-period, allows each sensor to study the variation of the monitored condition then to reduce its data collection according to two techniques, on-off transmission or adapting sensing frequency. The last phase, e.g. in-node, is applied at the CH level and aims to remove the redundancy among data collected by neighboring nodes, based on in-network correlation or data clustering techniques, before sending the data to the sink. We conducted simulations on real sensor data in order to evaluate the efficiency of our mechanism, in terms of several performance metrics, compared to other exiting techniques.

Quality of Service Improvement with Optimal Software-Defined Networking Controller and Control Plane Clustering

The controller is indispensable in software-defined networking (SDN). With several features, controllers monitor the network and respond promptly to dynamic changes. Their performance affects the quality-of-service (QoS) in SDN. Every controller supports a set of features. However, the support of the features may be more prominent in one controller. Moreover, a single controller leads to performance, single-point-of-failure (SPOF), and scalability problems. To overcome this, a controller with an optimum feature set must be available for SDN. Furthermore, a cluster of optimum feature set controllers will overcome an SPOF and improve the QoS in SDN. Herein, leveraging an analytical network process (ANP), we rank SDN controllers regarding their supporting features and create a hierarchical control plane based cluster (HCPC) of the highly ranked controller computed using the ANP, evaluating their performance for the OS3E topology. The results demonstrated in Mininet reveal that a HCPC environment with an optimum controller achieves an improved QoS. Moreover, the experimental results validated in Mininet show that our proposed approach surpasses the existing distributed controller clustering (DCC) schemes in terms of several performance metrics i.e., delay, jitter, throughput, load balancing, scalability and CPU (central processing unit) utilization.

ANALYZING THE PERFORMANCE OF PURE LATERATION IN INDOOR ENVIRONMENTS WITH VARIOUS PERFORMANCE METRICS

Nowadays, determining the location of the users and devices in indoor buildings is promising research topic. Accurate position determination of the users for indoor environments is used for numerous applications such as public safety, supermarkets, health care applications, travelling, social networks and tourism. However, global positioning systems created for outdoor localizations cannot be used for indoor positioning systems (IPS) because detecting the exact position of a target is an issue for IPS. For indoor environments, there are several positioning algorithms such as lateration, fingerprinting, dead reckoning etc. Lateration is low cost and easy to deploy when compared to other existing algorithms. Therefore, in this study, received signal strength based pure lateration that uses synthetic data generated from MATLAB is proposed. The performance of pure lateration is investigated in terms of several performance metrics such as effect of varying number of the access points (AP), varying dimensions of the measurement area, varying Gaussian Noise power and varying number of test points in the field. The simulation of the pure lateration algorithm is conducted in MATLAB. The effect of the performance metrics are investigated and discussed in details. According to the results, accuracy performance of lateration is increased when the number of APs increase in the area, however this will bring some hardware costs. In addition, when the number of test points increases in the field, in other words the step size between two test points decreases in the field the error performance of lateration is also enhanced however, this will also cause to computational costs. Finally, enlarging the measurement area causes to decrease the accuracy performance of lateration as expected. The main purpose of this study is to obtain the optimum conditions for lateration to provide a solution for real time applications. For future work, the real time implementations of this study are performed and to improve the accuracy performance, it is aimed to use a curve fitting idea to the measured values.

Improving Electric Energy Consumption Prediction Using CNN and Bi-LSTM

The electric energy consumption prediction (EECP) is an essential and complex task in intelligent power management system. EECP plays a significant role in drawing up a national energy development policy. Therefore, this study proposes an Electric Energy Consumption Prediction model utilizing the combination of Convolutional Neural Network (CNN) and Bi-directional Long Short-Term Memory (Bi-LSTM) that is named EECP-CBL model to predict electric energy consumption. In this framework, two CNNs in the first module extract the important information from several variables in the individual household electric power consumption (IHEPC) dataset. Then, Bi-LSTM module with two Bi-LSTM layers uses the above information as well as the trends of time series in two directions including the forward and backward states to make predictions. The obtained values in the Bi-LSTM module will be passed to the last module that consists of two fully connected layers for finally predicting the electric energy consumption in the future. The experiments were conducted to compare the prediction performances of the proposed model and the state-of-the-art models for the IHEPC dataset with several variants. The experimental results indicate that EECP-CBL framework outperforms the state-of-the-art approaches in terms of several performance metrics for electric energy consumption prediction on several variations of IHEPC dataset in real-time, short-term, medium-term and long-term timespans.

Bacteria Interactive Cost and Balanced-Compromised Approach to Clustering and Transmission Boundary-Range Cognitive Routing In Mobile Heterogeneous Wireless Sensor Networks.

The improvement of stable, energy-efficient mobile-based clustering and routing protocols in wireless sensor networks (WSNs) has become indispensable so as to develop large-scale, versitale, and adaptive applications. Data is gathered more efficiently and the total path length is shortened optimally by means of mobile sink (MS). Two algorithms as bacterial interaction based cluster head (CH) selection and energy and transmission boundary range cognitive routing algorithm with novel approach for heterogeneous mobile networks are proposed in this study. The more reliable and powerful CH selection is made with the greedy approach that is based on the interaction fitness value, energy node degree, and distance to adjacent nodes in a compromised manner. The best trajectories, thanks to intersection edge points of the visited CHs, are obtained in the proposed routing algorithm. In this way, the MS entry to transmission range boundaries of the CH has been a sufficient strategy to collect information. As in energy model, we adopt energy consumption costs of listening and sensing channel as well as transmit and receive costs. Comprehensive performance analyzes have been seriously carried out via the Matlab 2016a environment. We validate that the proposed scheme outperforms existing studies in terms of several performance metrics as simulations.

A Location-Aware Waypoint-Based Routing Protocol for Airborne DTNs in Search and Rescue Scenarios.

In this paper, we propose GeoSaW, a delay-tolerant routing protocol for Airborne Networks in Search and Rescue scenarios. The protocol exploits the geographical information of UAVs to make appropriate message forwarding decisions. More precisely, the information about the future UAV’s motion path is exploited to select the best UAV carrying the message towards the destination. Simulation results show that the proposed solution outperforms the classic DTN routing protocols in terms of several performance metrics.

Imbalanced classification of mental workload using a cost-sensitive majority weighted minority oversampling strategy

Identifying the temporal variations in mental workload level (MWL) is crucial for enhancing the safety of human---machine system operations, especially when there is cognitive overload or inattention of human operator. This paper proposed a cost-sensitive majority weighted minority oversampling strategy to address the imbalanced MWL data classification problem. Both the inter-class and intra-class imbalance problems are considered. For the former, imbalance ratio is defined to determine the number of the synthetic samples in the minority class. The latter problem is addressed by assigning different weights to borderline samples in the minority class based on the distance and density meaures of the sample distribution. Furthermore, multi-label classifier is designed based on an ensemble of binary classifiers. The results of analyzing 21 imbalanced UCI multi-class datasets showed that the proposed approach can effectively cope with the imbalanced classification problem in terms of several performance metrics including geometric mean (G-mean) and average accuracy (ACC). Moreover, the proposed approach was applied to the analysis of the EEG data of eight experimental participants subject to fluctuating levels of mental workload. The comparative results showed that the proposed method provides a competing alternative to several existing imbalanced learning algorithms and significantly outperforms the basic/referential method that ignores the imbalance nature of the dataset.

Structural failure classification for reinforced concrete buildings using trained neural network based multi-objective genetic algorithm

Structural design has an imperative role in deciding the failure possibility of a Reinforced Concrete (RC) structure. Recent research works achieved the goal of predicting the structural failure of the RC structure with the assistance of machine learning techniques. Previously, the Artificial Neural Network (ANN) has been trained supported by Particle Swarm Optimization (PSO) to classify RC structures with reasonable accuracy. Though, keeping in mind the sensitivity in predicting the structural failure, more accurate models are still absent in the context of Machine Learning. Since the efficiency of multiobjective optimization over single objective optimization techniques is well established. Thus, the motivation of the current work is to employ a Multi-objective Genetic Algorithm (MOGA) to train the Neural Network (NN) based model. In the present work, the NN has been trained with MOGA to minimize the Root Mean Squared Error (RMSE) and Maximum Error (ME) toward optimizing the weight vector of the NN. The model has been tested by using a dataset consisting of 150 RC structure buildings. The proposed NN-MOGA based model has been compared with Multi-layer perceptron-feed-forward network (MLPFFN) and NN-PSO based models in terms of several performance metrics. Experimental results suggested that the NN-MOGA has outperformed other existing well known classifiers with a reasonable improvement over them. Meanwhile, the proposed NN-MOGA achieved the superior accuracy of 93.33% and F-measure of 94.44%, which is superior to the other classifiers in the present study.

Robust RST control design based on Multi-Objective Particle Swarm Optimization approach

In this paper, a novel method for the digital two-Degrees-Of-Freedom (2DOF) controller design, called canonical RST structure, is proposed and successfully implemented based on a Multi-Objective Particle Swarm Optimization (MOPSO) approach. This is a polynomial control structure allowing independently the regulation and the tracking of discrete-time systems. An application to the variable speed control of an electrical DC Drive is investigated. The RST design and tuning problem is formulated as a multi-objective optimization problem. The proposed MOPSO algorithm which is based on the Pareto dominance is used to identify the non-dominated solutions. This approach used the leader selection strategy that is called a geographically-based system. In addition, the adaptive grid method is used to produce well-distributed Pareto fronts in the multi-objective formalism. The well known NSGA-II and the proposed MOPSO algorithms are evaluated and compared with each other in terms of several performance metrics in order to show the superiority and the effectiveness of the proposed method. Simulation results demonstrate the advantages of the MOPSO-tuned RST control structure in terms of performance and robustness.

Identification of Mental Workload Using Imbalanced EEG Data and DySMOTE-based Neural Network Approach

Identifying the temporal changes of mental workload level (MWL) is crucial for enhancing the safety of human-machine (HM) system operations especially when human operators suffer from cognitive overload and inattention. In this paper, we oversampled the EEG data in minority class to dynamically learn a multilayer perceptron (MLP) model based on a dynamical SMOTE (DySMOTE) approach in order to classify the MWL. The proposed approach consists of data sampling and a dynamical selection strategy, in which the probability of each sample being selected to update the weights and thresholds of the MLP is estimated in each epoch to derive an accurate classifier model. The DySMOTE approach was evaluated on the measured EEG data from eight subjects. The results showed that the proposed method outperforms existing methods in terms of several performance metrics including geometric mean (G-mean) and classification accuracy (or correct classification rate) of individual classes of the MWL.