Terms Of Classification Accuracy Research Articles

A comparative analysis of haemoglobin variants using machine learning algorithms

In medical sciences, to ascertain the origin of a sickness, professionals utilize their expertise and knowledge to analyze a person's symptoms and indications. These symptoms (indicators) are threshold values that health specialists use to determine the cause of the illness by comparing a specific proportion of measurements to where a healthy population would fall. Consequently, diagnostic mistakes occur as a result of inaccuracy and imprecision. This study utilizes machine learning to categorize haemoglobin variations. Specifically, the data set used in this study includes 752 complete blood count laboratory analyses of adult patients aged eighteen and above obtained from Lagos State University Teaching Hospital (LASUTH). Multiple machine learning methods were utilized for classification from which five of the methods employed were examined and assessed. Comparative analysis was done using the five algorithms (K-Nearest Neighbors (KNN), Support Vector Machine (SVM), Decision Tree (DT), Random Forest (RF), and Naive Bayes (NB)). Contrary to work done by previous researchers, it was observed that the SVM model showed the best classification accuracy of 94.7%, with an F1-score of 94.5%, precision of 94.8%, recall of 94.7%, specificity of 97.3%, and area under curve (AUC) of 99.0%. Among the other models considered, the RF model gave the least accuracy result of 87.4%. The study shows that the support vector machine algorithm outperforms the other classifiers in terms of accuracy when predicting haemoglobin variants given the haematological parameters.

Residential Appliance Identification Using 1-D Convolutional Neural Network Based on Multiscale Sinusoidal Initializers

Appliance identification and classification are primary requirement for successful deployment of demand side management in smart grid. Traditionally, load identification is carried out after extracting handcrafted features from electrical signals followed by classification algorithms, which is laborious and time consuming. To nullify the abovementioned shortcomings, this article proposes a 1-D convolutional neural network (CNN) with sinusoidal kernel initializers for classification of domestic appliances operating individually or in combination with other devices. For efficient operation, 15 primary and 225 multiscale sinusoidal kernels (MSK) are formed and included in the CNN architecture for extracting important discriminative features from the current signals. Classification accuracy of the proposed CNN module, named as MSK-CNN, is further improved by adding a new nonlinear activation function (AF), SL-ReLU, having an adjustable logarithmic function along with softsign and ReLU functions. The efficacy of MSK-CNN is tested on a practical dataset, created by the authors, consists current signals of different real life residential appliances. It is showed that MSK-CNN can successfully distinguish 18 single and 12 appliance combinations, and can attain an accuracy of 98.61%. The experimental outcomes reveal superiority of the MSK-CNN over classical CNN modules with other kernel initializers and AFs in terms of classification accuracy and training convergence.

Brain Tumor Detection Using CNN

Abstract: Brain tumors being one of the most fatal illnesses for people across all ages, they make up to 90% of tumors that principally affect the central nervous system. To ensure recovery and highest possible life expectancy for the patient doctors must ensure precise diagnosis and properly plan out treatment courses. Manual screening of tumors of the brain is highly complex. Having a tendency to being misdiagnosed owing to Byzantine structure of the brain and varying characteristics of tumors depending on their location and other factors. With the rise of computer sciences, there are automatic detection and classification solutions available that can outperform manual diagnosis and classification in terms of accuracy using machine learning techniques, Medical Professionals around the globe could use such systems. Comprehensive experiments were conducted by us on the selected datasets and it’s shown that the proposed model is in a position to realize competitive results.

Radio Frequency Fingerprint extraction based on Multiscale Approximate Entropy

Radio Frequency (RF) fingerprinting is the technology of recognizing the transmitter using the non-linear characteristics of an intercepted RF signal. The underlying inevitable impairments of the hardware chain in transmitters are used as unique RF signatures for the distinct identification of various radios. The technique can be applied to distinguish not only between the radio of different make but also between the radios of the same make and type. In this paper, we propose a novel RF fingerprinting method, based on Multi-Scale Approximate Entropy (MSAE) which utilizes the steady-state section of the RF signal, extracted through the Higuchi Fractal Dimension (HFD) method. The MSAE feature extraction method is validated using real-world data-set for Very High Frequency (VHF) radios. The proposed method uses MSAE features which are subsequently fed to Machine Learning (ML) algorithms for classification accuracy comparison. In terms of classification accuracy, the proposed MSAE features outperforms some of the existing steady-state methods, especially at low SNR.

MixNN: A Design for Protecting Deep Learning Models.

In this paper, we propose a novel design, called MixNN, for protecting deep learning model structure and parameters since the model consists of several layers and each layer contains its own structure and parameters. The layers in a deep learning model of MixNN are fully decentralized. It hides communication address, layer parameters and operations, and forward as well as backward message flows among non-adjacent layers using the ideas from mix networks. MixNN has the following advantages: (i) an adversary cannot fully control all layers of a model, including the structure and parameters; (ii) even some layers may collude but they cannot tamper with other honest layers; (iii) model privacy is preserved in the training phase. We provide detailed descriptions for deployment. In one classification experiment, we compared a neural network deployed in a virtual machine with the same one using the MixNN design on the AWS EC2. The result shows that our MixNN retains less than 0.001 difference in terms of classification accuracy, while the whole running time of MixNN is about 7.5 times slower than the one running on a single virtual machine.

The real-time and stack fusion enhanced dual-channel network with attention modules for fast hyperspectral image classification

Hyperspectral image (HSI) classification with convolutional neural network (CNN) exhibiting excellent performance, but the improvement of computationally efficient is yet a challenging research task. Therefore, two fusion strategies and two attention modules have been introduced to enhance for a dual-channel structure, which consists of 2-D convolution branch for pure spatial features and 3-D convolution branch for joint spectral-spatial features. The real-time fusion is applied to each convolution layer between two branches, which enhanced information interaction within the network. Meanwhile, the stacked fusion strategy with the adaptive weights is introduced to adjust the influence of the two branches on the final classification results. Furthermore, the spectral-spatial reconstruction attention module (SSRAM) and adaptive channel attention module (ACAM) are designed to enhance features representative capability for 3-D branch and 2-D branch, respectively. All approaches introduced above allow us to construct a dual-channel network consisting of small convolution kernels, which achieve HSI classification accurately and efficiently. Experimental results on three public HSI datasets show that the proposed network outperforms several state-of-the-art models in terms of classification accuracy and computational efficiency.

An Enhanced Spectral Fusion 3D CNN Model for Hyperspectral Image Classification

With the continuous development of hyperspectral image technology and deep learning methods in recent years, an increasing number of hyperspectral image classification models have been proposed. However, due to the numerous spectral dimensions of hyperspectral images, most classification models suffer from issues such as breaking spectral continuity and poor learning of spectral information. In this paper, we propose a new classification model called the enhanced spectral fusion network (ESFNet), which contains two parts: an optimized multi-scale fused spectral attention module (FsSE) and a 3D convolutional neural network (3D CNN) based on the fusion of different spectral strides (SSFCNN). Specifically, after sampling the hyperspectral images, our model first implements the weighting of the spectral information through the FsSE module to obtain spectral data with a higher degree of information richness. Then, the weighted spectral data are fed into the SSFCNN to realize the effective learning of spectral features. The new model can maximize the retention of spectral continuity and enhance the spectral information while being able to better utilize the enhanced information to improve the model’s ability to learn hyperspectral image features, thus improving the classification accuracy of the model. Experiment results on the Indian Pines and Pavia University datasets demonstrated that our method outperforms other relevant baselines in terms of classification accuracy and generalization performance.

Rapid gas detection for electronic noses using optimal measurement time search and multi-sensor energy change based feature extraction

Rapid detection of gases using electronic noses is essential and challenging owing to the associated requirements of high accuracy recognition and rapid measurement. In this study, we propose a strategy based on performing feature extraction followed by searching for the appropriate measurement time to improve the measurement speed. To implement our strategy and further improve classification accuracy, we propose a novel feature extraction algorithm based on the energy change of multi-sensor. The superiority of the proposed approach was verified by six classification models on 15 electronic nose datasets. The experimental results show that it outperforms other typical algorithms in terms of classification accuracy. Moreover, the proposed approach achieved reliable estimates using only the first 23.68% of the measurement data on average. The findings indicate the effectiveness of the proposed method in improving the accuracy and measurement speed, which is expected to contribute to the future design of rapid detection systems.

Efficient spatio-temporal feature clustering for large event-based datasets

Event-based cameras encode changes in a visual scene with high temporal precision and low power consumption, generating millions of events per second in the process. Current event-based processing algorithms do not scale well in terms of runtime and computational resources when applied to a large amount of data. This problem is further exacerbated by the development of high spatial resolution vision sensors. We introduce a fast and computationally efficient clustering algorithm that is particularly designed for dealing with large event-based datasets. The approach is based on the expectation-maximization (EM) algorithm and relies on a stochastic approximation of the E-step over a truncated space to reduce the computational burden and speed up the learning process. We evaluate the quality, complexity, and stability of the clustering algorithm on a variety of large event-based datasets, and then validate our approach with a classification task. The proposed algorithm is significantly faster than standard k-means and reduces computational demands by two to three orders of magnitude while being more stable, interpretable, and close to the state of the art in terms of classification accuracy.

Matched Filter Interpretation of CNN Classifiers with Application to HAR.

Time series classification is an active research topic due to its wide range of applications and the proliferation of sensory data. Convolutional neural networks (CNNs) are ubiquitous in modern machine learning (ML) models. In this work, we present a matched filter (MF) interpretation of CNN classifiers accompanied by an experimental proof of concept using a carefully developed synthetic dataset. We exploit this interpretation to develop an MF CNN model for time series classification comprising a stack of a Conv1D layer followed by a GlobalMaxPooling layer acting as a typical MF for automated feature extraction and a fully connected layer with softmax activation for computing class probabilities. The presented interpretation enables developing superlight highly accurate classifier models that meet the tight requirements of edge inference. Edge inference is emerging research that addresses the latency, availability, privacy, and connectivity concerns of the commonly deployed cloud inference. The MF-based CNN model has been applied to the sensor-based human activity recognition (HAR) problem due to its significant importance in a broad range of applications. The UCI-HAR, WISDM-AR, and MotionSense datasets are used for model training and testing. The proposed classifier is tested and benchmarked on an android smartphone with average accuracy and scores of 98% and 97%, respectively, which outperforms state-of-the-art HAR methods in terms of classification accuracy and run-time performance. The proposed model size is less than 150 KB, and the average inference time is less than 1 ms. The presented interpretation helps develop a better understanding of CNN operation and decision mechanisms. The proposed model is distinguished from related work by jointly featuring interpretability, high accuracy, and low computational cost, enabling its ready deployment on a wide set of mobile devices for a broad range of applications.

Intramuscular EMG classifier for detecting myopathy and neuropathy

This article presents an automatic diagnostic system to classify intramuscular electromyography (iEMG) signals, thereby detecting neuromuscular disorders. To this end, we tailored the center symmetric local binary pattern (CSLBP) to analyze one-dimensional ( 1 -D) signals. In this approach, the 1 -D CSLBP feature extracted from a decimated iEMG signal is fed to a combination of classifiers, which in turn assigns a set of labels to the signal, and ultimately the signal category is determined by the Boyer-Moore majority voting (BMMV) algorithm. The proposed framework was investigated with a benchmark iEMG dataset that contains signals recorded from three different muscles: biceps brachii (BB), deltoideus (DE), and vastus medialis (VM). In a repeated 10 -fold cross-validation, CSLBP-Combined-Classifiers-BMMV (CSLBP-CC-BMMV) achieved an average classification accuracy of 92.80 %, 94.25 %, and 93.71 % for the iEMG signals recorded from BB, DE, and VM muscle, respectively. Interestingly, the performance of CSLBP-CC-BMMV surpassed the other published approaches and ensemble learning methods that are akin to our scheme in terms of classification accuracy and computational time.

Real-time terahertz characterization for composite delamination using a lightweight CPU adaptive network

Delamination defects, as common damages occurring in fiber reinforced composite laminates, are often difficult to be detected due to their invisibility and concealment. Terahertz (THz) technique, as a promising alternative to conventional nondestructive testing (NDT) approaches, has emerged great potentials in quantitative characterization of delamination defects in composites. However, since THz signals are susceptible to confounding interference during THz inspection, such as the noise, interlayer reflection, dispersion and overlap, the accuracy and speed of defect localization and identification are usually contradictory. Generally, to improve the accuracy, many complex manual signal processing methods can be applied, which however will inevitably decrease the inference speed for delamination inspection and cannot fulfill the automatic and real-time analysis requirements in practical THz applications. Here, an intelligent THz 3D characterization system based on the lightweight deep learning (DL) model is proposed to realize the automatic and real-time characterization of hidden delamination defects in composites with high accuracy and low CPU latency. The core aims to specifically design a lightweight and end-to-end CPU adaptive network to obtain the optimal balance between the accuracy and speed. A series of experiments are implemented to validate the superior comprehensive performance of the system in terms of classification accuracy and inference speed on low-cost CPU-based platform. Compared with other methods, the proposed system can be not only able to obtain the superior accuracy, but also to obtain low CPU latency. Therefore, our work provides a novel and general paradigm for DL-based THz intelligent 3D characterization, which will promote the deployment of the system for the automatic and real-time THz characterization in industrial applications.

An Efficient Hybrid Feature Selection Method Using the Artificial Immune Algorithm for High-Dimensional Data.

Feature selection provides the optimal subset of features for data mining models. However, current feature selection methods for high-dimensional data also require a better balance between feature subset quality and computational cost. In this paper, an efficient hybrid feature selection method (HFIA) based on artificial immune algorithm optimization is proposed to solve the feature selection problem of high-dimensional data. The algorithm combines filter algorithms and improves clone selection algorithms to explore the feature space of high-dimensional data. According to the target requirements of feature selection, combined with biological research results, this method introduces the lethal mutation mechanism and the Cauchy operator to improve the search performance of the algorithm. Moreover, the adaptive adjustment factor is introduced in the mutation and update phases of the algorithm. The effective combination of these mechanisms enables the algorithm to obtain a better search ability and lower computational costs. Experimental comparisons with 19 state-of-the-art feature selection methods are conducted on 25 high-dimensional benchmark datasets. The results show that the feature reduction rate for all datasets is above 99%, and the performance improvement for the classifier is between 5% and 48.33%. Compared with the five classical filtering feature selection methods, the computational cost of HFIA is lower than the two of them, and it is far better than these five algorithms in terms of the feature reduction rate and classification accuracy improvement. Compared with the 14 hybrid feature selection methods reported in the latest literature, the average winning rates in terms of classification accuracy, feature reduction rate, and computational cost are 85.83%, 88.33%, and 96.67%, respectively.

Early detection and classification of liver diseases in ultrasound images using hybrid machine learning techniques

Liver disease is a form of sickness that starts in the ovaries and is particularly dangerous for women. As a result, aberrant cells develop that have the potential to spread to other parts of the body. Liver disease is a sort of risky improvement that impacts ovaries in ladies, and is hard to perceive at early phase because of which it stays as one of the guideline wellsprings of illness end. Unquestionable confirmation of intrinsic and typical parts is immense in making novel frameworks to perceive and ruin danger. Here, using hybrid machine learning approaches, we propose an early detection and classification of liver disease (EDCLD) in ultrasound images. This study is aimed at detecting and evaluating liver disease, which may be benign or malignant depending on the cancer type. To begin with, we use an improved ant swarm optimization algorithm for both high and low level feature extraction models. For liver disease classification, we used hybrid swallow swarm intelligent-deep neural network (HSSI-DNN) classifier to identify the disease stage. By using training and testing liver datasets, we are improving the accuracy of liver disease detection with several stages. Second, we present a search optimization for whales algorithm for cascaded and segmentation as a single fusion feature capable of representing both the texture patterns and semantic background distributed in the image. Third, the dividend yields are commitment to the HSSI-DNN classifier used to detect the “Normal” and “Diseases” from input image. The proposed HSSI-DNN classifier performance outperformed the existing classifier in terms of accuracy (0.98), specificity (0.98), and sensitivity (1). The proposed EDCLD technique implemented in MATLAB.

Opposition-based sine cosine optimizer utilizing refraction learning and variable neighborhood search for feature selection.

This paper proposes new improved binary versions of the Sine Cosine Algorithm (SCA) for the Feature Selection (FS) problem. FS is an essential machine learning and data mining task of choosing a subset of highly discriminating features from noisy, irrelevant, high-dimensional, and redundant features to best represent a dataset. SCA is a recent metaheuristic algorithm established to emulate a model based on sine and cosine trigonometric functions. It was initially proposed to tackle problems in the continuous domain. The SCA has been modified to Binary SCA (BSCA) to deal with the binary domain of the FS problem. To improve the performance of BSCA, three accumulative improved variations are proposed (i.e., IBSCA1, IBSCA2, and IBSCA3) where the last version has the best performance. IBSCA1 employs Opposition Based Learning (OBL) to help ensure a diverse population of candidate solutions. IBSCA2 improves IBSCA1 by adding Variable Neighborhood Search (VNS) and Laplace distribution to support several mutation methods. IBSCA3 improves IBSCA2 by optimizing the best candidate solution using Refraction Learning (RL), a novel OBL approach based on light refraction. For performance evaluation, 19 real-wold datasets, including a COVID-19 dataset, were selected with different numbers of features, classes, and instances. Three performance measurements have been used to test the IBSCA versions: classification accuracy, number of features, and fitness values. Furthermore, the performance of the last variation of IBSCA3 is compared against 28 existing popular algorithms. Interestingly, IBCSA3 outperformed almost all comparative methods in terms of classification accuracy and fitness values. At the same time, it was ranked 15 out of 19 in terms of number of features. The overall simulation and statistical results indicate that IBSCA3 performs better than the other algorithms.

Multi-Objective Particle Swarm Optimization Based Preprocessing of Multi-Class Extremely Imbalanced Datasets

Today’s datasets are usually very large with many features and making analysis on such datasets is really a tedious task. Especially when performing classification, selecting attributes that are salient for the process is a brainstorming task. It is more difficult when there are many class labels for the target class attribute and hence many researchers have introduced methods to select features for performing classification on multi-class attributes. The process becomes more tedious when the attribute values are imbalanced for which researchers have contributed many methods. But, there is no sufficient research to handle extreme imbalance and feature selection together and hence this paper aims to bridge this gap. Here Particle Swarm Optimization (PSO), an efficient evolutionary algorithm is used to handle imbalanced dataset and feature selection process is also enhanced with the required functionalities. First, Multi-objective Particle Swarm Optimization is used to transform the imbalanced datasets into balanced one and then another version of Multi-objective Particle Swarm Optimization is used to select the significant features. The proposed methodology is applied on eight multi-class extremely imbalanced datasets and the experimental results are found to be better than other existing methods in terms of classification accuracy, G mean, F measure. The results validated by using Friedman test also confirm that the proposed methodology effectively balances the dataset with less number of features than other methods.

A novel CapsNet neural network based on MobileNetV2 structure for robot image classification.

Image classification indicates that it classifies the images into a certain category according to the information in the image. Therefore, extracting image feature information is an important research content in image classification. Traditional image classification mainly uses machine learning methods to extract features. With the continuous development of deep learning, various deep learning algorithms are gradually applied to image classification. However, traditional deep learning-based image classification methods have low classification efficiency and long convergence time. The training networks are prone to over-fitting. In this paper, we present a novel CapsNet neural network based on the MobileNetV2 structure for robot image classification. Aiming at the problem that the lightweight network will sacrifice classification accuracy, the MobileNetV2 is taken as the base network architecture. CapsNet is improved by optimizing the dynamic routing algorithm to generate the feature graph. The attention module is introduced to increase the weight of the saliency feature graph learned by the convolutional layer to improve its classification accuracy. The parallel input of spatial information and channel information reduces the computation and complexity of network. Finally, the experiments are carried out in CIFAR-100 dataset. The results show that the proposed model is superior to other robot image classification models in terms of classification accuracy and robustness.

Multiplicative Long Short-Term Memory with Improved Mayfly Optimization for LULC Classification

Land Use and Land Cover (LULC) monitoring is crucial for global transformation, sustainable land control, urban planning, urban growth prediction, and the establishment of climate regulations for long-term development. Remote sensing images have become increasingly important in many environmental planning and land use surveys in recent times. LULC is evaluated in this research using the Sat 4, Sat 6, and Eurosat datasets. Various spectral feature bands are involved, but unexpectedly little consideration has been given to these characteristics in deep learning models. Due to the wide availability of RGB models in computer vision, this research mainly utilized RGB bands. Once the pre-processing is carried out for the images of the selected dataset, the hybrid feature extraction is performed using Haralick texture features, an oriented gradient histogram, a local Gabor binary pattern histogram sequence, and Harris Corner Detection to extract features from the images. After that, the Improved Mayfly Optimization (IMO) method is used to choose the optimal features. IMO-based feature selection algorithms have several advantages that include features such as a high learning rate and computational efficiency. After obtaining the optimal feature selection, the LULC classes are classified using a multi-class classifier known as the Multiplicative Long Short-Term Memory (mLSTM) network. The main functionality of the multiplicative LSTM classifier is to recall appropriate information for a comprehensive duration. In order to accomplish an improved result in LULC classification, a higher amount of remote sensing data should be processed. So, the simulation outcomes demonstrated that the proposed IMO-mLSTM efficiently classifies the LULC classes in terms of classification accuracy, recall, and precision. When compared with ConvNet and Alexnet, the proposed IMO-mLSTM method accomplished accuracies of 99.99% on Sat 4, 99.98% on Sat 6, and 98.52% on the Eurosat datasets.

Improving the Performance and Explainability of Indoor Human Activity Recognition in the Internet of Things Environment

Traditional indoor human activity recognition (HAR) has been defined as a time-series data classification problem and requires feature extraction. The current indoor HAR systems still lack transparent, interpretable, and explainable approaches that can generate human-understandable information. This paper proposes a new approach, called Human Activity Recognition on Signal Images (HARSI), which defines the HAR problem as an image classification problem to improve both explainability and recognition accuracy. The proposed HARSI method collects sensor data from the Internet of Things (IoT) environment and transforms the raw signal data into some visual understandable images to take advantage of the strengths of convolutional neural networks (CNNs) in handling image data. This study focuses on the recognition of symmetric human activities, including walking, jogging, moving downstairs, moving upstairs, standing, and sitting. The experimental results carried out on a real-world dataset showed that a significant improvement (13.72%) was achieved by the proposed HARSI model compared to the traditional machine learning models. The results also showed that our method (98%) outperformed the state-of-the-art methods (90.94%) in terms of classification accuracy.

A New Algorithm for Cancer Biomarker Gene Detection Using Harris Hawks Optimization

This paper presents two novel swarm intelligence algorithms for gene selection, HHO-SVM and HHO-KNN. Both of these algorithms are based on Harris Hawks Optimization (HHO), one in conjunction with support vector machines (SVM) and the other in conjunction with k-nearest neighbors (k-NN). In both algorithms, the goal is to determine a small gene subset that can be used to classify samples with a high degree of accuracy. The proposed algorithms are divided into two phases. To obtain an accurate gene set and to deal with the challenge of high-dimensional data, the redundancy analysis and relevance calculation are conducted in the first phase. To solve the gene selection problem, the second phase applies SVM and k-NN with leave-one-out cross-validation. A performance evaluation was performed on six microarray data sets using the two proposed algorithms. A comparison of the two proposed algorithms with several known algorithms indicates that both of them perform quite well in terms of classification accuracy and the number of selected genes.