Feature Extraction Tool Research Articles

Hierarchical diversity entropy for the early fault diagnosis of rolling bearing

Intelligent fault diagnosis provides great convenience for the prognostic and health management of the rotating machinery. Recently, the multiscale diversity entropy has been proven to be a promising feature extraction tool for the intelligent fault diagnosis. Compared with the existing entropy methods, the multiscale diversity entropy has advantages of high consistency, strong robustness, and high calculation efficiency. However, the multiscale diversity entropy encounters the challenge to extract features from early fault signals with weak fault symptoms and strong noise. This can be attributed to the multiscale diversity entropy that only concerns the fault information embedded in the low frequency, which ignores the information hidden in the high frequency. To address this defect, the hierarchical diversity entropy (HDE) is proposed, which can synchronously extract fault information hidden in both high and low frequencies. Based on HDE and random forest, a novel intelligent fault diagnosis frame has been proposed. The effectiveness of the proposed method has been evaluated through simulated and experimental bearing signals. The results show that the proposed HDE has the best feature extraction ability compared with multiscale sample entropy, multiscale permutation entropy, multiscale fuzzy entropy, and multiscale diversity entropy.

PSSMCOOL: a comprehensive R package for generating evolutionary-based descriptors of protein sequences from PSSM profiles.

Position-specific scoring matrix (PSSM), also called profile, is broadly used for representing the evolutionary history of a given protein sequence. Several investigations reported that the PSSM-based feature descriptors can improve the prediction of various protein attributes such as interaction, function, subcellular localization, secondary structure, disorder regions, and accessible surface area. While plenty of algorithms have been suggested for extracting evolutionary features from PSSM in recent years, there is not any integrated standalone tool for providing these descriptors. Here, we introduce PSSMCOOL, a flexible comprehensive R package that generates 38 PSSM-based feature vectors. To our best knowledge, PSSMCOOL is the first PSSM-based feature extraction tool implemented in R. With the growing demand for exploiting machine-learning algorithms in computational biology, this package would be a practical tool for machine-learning predictions.

Optical fibre based quench detection in HTS applications using machine learning classifiers

A Mach-Zehnder Interferometer (MZI) based optical fibre sensing technique, developed and patented by EPFL, is an efficient and economical way to detect hotspots in High Temperature Superconductor (HTS) applications. Due to the MZI sensitivity being a composite of strain sensitive and temperature sensitive contributions, the MZI gives an instantaneous response to a quench (within 10 ms), because of the quick strain transfer to the optical fibre. However, the MZI output signal can also manifest the environmental noise caused by mechanical vibrations, bubbling in the cryostat and temperature variations, along with the response to the quench. This presents the problems of false alarms and indiscernible response to a quench. Discrete wavelet transform (DWT) has been proven to be a useful tool for feature extraction in different fields requiring signal categorization and hence holds the potential to enable quench recognition in the MZI output. This paper proposes an effective approach of performing DWT based feature extraction on experimental data and subsequently using the extracted features for the MZI response classification using two machine learning based classification techniques: k-nearest neighbours (KNN) and Artificial Neural Network (ANN). For this manuscript, experiments were performed using MZI for quench detection in an HTS tape. Feature extraction was then implemented on these experimental measurements using discrete wavelet coefficients extracted at different decomposition levels from the MZI output; these features were then used to train the KNN and ANN models for identifying quench in the MZI signal. This method could be a valuable supplement to the MZI technique by enabling the development of a real time application that can process the MZI output data as well as eliminate the occurrences of false alarms; thereby facilitating reliable quench detection. With this development, the MZI technique would become an even more attractive solution for the health monitoring of HTS applications.

FEPS: A Tool for Feature Extraction from Protein Sequence.

Machine learning has become one of the most popular choices for developing computational approaches in protein structural bioinformatics. The ability to extract features from protein sequence/structure often becomes one of the crucial steps for the development of machine learning-based approaches. Over the years, various sequence, structural, and physicochemical descriptors have been developed for proteins and these descriptors have been used to predict/solve various bioinformatics problems. Hence, several feature extraction tools have been developed over the years to help researchers to generate numeric features from protein sequences. Most of these tools have some limitations regarding the number of sequences they can handle and the subsequent preprocessing that is required for the generated features before they can be fed to machine learning methods. Here, we present Feature Extraction from Protein Sequences (FEPS), a toolkit for feature extraction. FEPS is a versatile software package for generating various descriptors from protein sequences and can handle several sequences: the number of which is limited only by the computational resources. In addition, the features extracted from FEPS do not require subsequent processing and are ready to be fed to the machine learning techniques as it provides various output formats as well as the ability to concatenate these generated features. FEPS is made freely available via an online web server as well as a stand-alone toolkit. FEPS, a comprehensive toolkit for feature extraction, will help spur the development of machine learning-based models for various bioinformatics problems.

Advances in Hyperspectral Image Classification Based on Convolutional Neural Networks: A Review

Hyperspectral image (HSI) classification has been one of the most important tasks in the remote sensing community over the last few decades. Due to the presence of highly correlated bands and limited training samples in HSI, discriminative feature extraction was challenging for traditional machine learning methods. Recently, deep learning based methods have been recognized as powerful feature extraction tool and have drawn a significant amount of attention in HSI classification. Among various deep learning models, convolutional neural networks (CNNs) have shown huge success and offered great potential to yield high performance in HSI classification. Motivated by this successful performance, this paper presents a systematic review of different CNN architectures for HSI classification and provides some future guidelines. To accomplish this, our study has taken a few important steps. First, we have focused on different CNN architectures, which are able to extract spectral, spatial, and joint spectral-spatial features. Then, many publications related to CNN based HSI classifications have been reviewed systematically. Further, a detailed comparative performance analysis has been presented between four CNN models namely 1D CNN, 2D CNN, 3D CNN, and feature fusion based CNN (FFCNN). Four benchmark HSI datasets have been used in our experiment for evaluating the performance. Finally, we concluded the paper with challenges on CNN based HSI classification and future guidelines that may help the researchers to work on HSI classification using CNN.

A Survey of Smart Home IoT Device Classification Using Machine Learning-Based Network Traffic Analysis

Smart home IoT devices lack proper security, raising safety and privacy concerns. One-size-fits-all network administration is ineffective because of the diverse QoS requirements of IoT devices. Device classification can improve IoT administration and security. It identifies vulnerable and rogue items and automates network administration by device type or function. Considering this, a promising research topic focusing on Machine Learning (ML)-based traffic analysis has emerged in order to demystify hidden patterns in IoT traffic and enable automatic device classification. This study analyzes these approaches to understand their potential and limitations. It starts by describing a generic workflow for IoT device classification. It then looks at the methods and solutions for each stage of the workflow. This mainly consists of i) an analysis of IoT traffic data acquisition methodologies and scenarios, as well as a classification of public datasets, ii) a literature evaluation of IoT traffic feature extraction, categorizing and comparing popular features, as well as describing open-source feature extraction tools, and iii) a comparison of ML approaches for IoT device classification and how they have been evaluated. The findings of the analysis are presented in taxonomies with statistics showing literature trends. This study also explores and suggests undiscovered or understudied research directions.

Time-Domain Protection of Superconducting Cables Based on Artificial Intelligence Classifiers

Fault detection and protection of Superconducting Cables (SCs) is considered a challenging task due to the effects of the quenching phenomenon of High Temperature Superconducting (HTS) tapes and the prospective magnitude of fault currents in presence of highly-resistive faults and converter-interfaced generation. This paper presents a novel, time-domain method for discriminative detection of faults in a power system incorporating SCs and high penetration of renewable energy sources. The proposed algorithms utilizes feature extraction tools based on Stationary Wavelet Transform (SWT), as well as artificial intelligence (AI) classifiers to discriminate between external and internal faults, and other network events. The performance of the proposed schemes has been validated in electromagnetic transient simulation environment using a verified model of SC. Simulation results revealed that the proposed algorithms can effectively and within short period of time discriminate internal faults occurring on SC, while remain stable to external faults and other disturbances. The suitability of the proposed algorithms for real-time implementation has been verified using software and hardware in the loop testing environment. To determine the best options for real-time deployment, two different artificial intelligence classifiers namely Artificial Neural Network (ANN) and Support Vector Machine (SVM) have been deployed. The extensive assessment of their performance revealed that the ANN classifier is advantageous in term of prediction speed.

A New Way of Airline Traffic Prediction Based on GCN-LSTM.

With the development of society and the improvement of people's material level, more and more people like to travel by airplane. If we can predict the passenger flow of an airline in advance, it can be used as an important decision-making basis for its flight route planning, crew scheduling planning and ticket price formulation in the process of management and operation. However, due to the high complexity of aviation network, the existing traffic prediction methods generally have the problem of low prediction accuracy. In order to overcome this problem, this paper makes full use of graph convolutional neural network and long—short memory network to construct a prediction system with short—term prediction ability. Specifically, this paper uses the graph convolutional neural network as a feature extraction tool to extract the key features of air traffic data, and solves the problem of long term and short term dependence between data through the long term memory network, then we build a high-precision air traffic prediction system based on it. Finally, we design a comparison experiment to compare the algorithm with the traditional algorithms. The results show that the algorithm we proposed in this paper has a higher accuracy in air flow prediction according to the lower loss function value.

Mapping the evolution of topics published by Education for Information: Interdisciplinary Journal of Information Studies

The need to map the evolution of trends in any field of activity arises when a large amount of data is available on that activity, thus making impossible a manual exploration of the data in order to understand how the field or the activity is evolving. Topic and trend mapping is a mature field with hundreds of publications on approaches, methods and tools for data collection, analysis, feature extraction and reduction, clustering and visualisation tools and algorithms. Our study aims to map the evolution of topics published by the journal Education for Information. Interdisciplinary Journal on Information Studies (EFI henceforth) which has been in existence since 1983, in order to understand how this journal has evolved and how it is positioned with regard to the field of Library and Information Science to which it belongs. Our study is part of the body of work on topic detection and text mining. Our results showed that the journal displayed a remarkable stability in its editorial policy over more than three decades. With the arrival of its third Editor in Chief in 2018, a shift towards more technologically oriented topics and to specialties from other fields are perceptible such as health information, data science and digital humanities.

Breast Cancer Prediction using Machine Learning

With the increase in occurrence of human diseases all over the world, it has created a wide number of opportunities for innovating such kind of disease prediction mechanisms. Out of which, Breast Cancer has grown in tremendous number with an increased amount in the past decade and this has gone increasing till now, & would continue to grow. Now moving further, there is a tremendous need for efficient text analytics tools and feature extraction tools to assist the work related in classifying, sharing and retrieving the information on human diseases in general and in the field of Breast Cancer also. In addition to that, the present study has been made with the objective to provide a complete analysis of different classifiers on Breast Cancer dataset, and to generate a new ensemble training method/model of Machine Learning Classification. Here, machine learning algorithms (such as Decision Tree , Logistic Regression ,Support Vector Machine, Random Forest). An Ensemble Learning model for Prediction is proposed to classify the results among different classifiers as mentioned in the above statement. At the end, the Voting Ensemble technique is implemented in order to find out the optimal classifier that predicts the Breast Cancer. The results have been computed on the basis of evaluation parameters such as Accuracy, Precision, Recall and Specificity. On the other hand, the confusion matrix is measured on the basis of evaluation parameters that generally provides more emphasised, predicted and actual instance of data. Whereas, Performance Evaluation for various machine learning algorithms is then computed. Results of this investigation says that, the Voting Ensemble superpowers other machine learning algorithms.

Microfluidic droplet detection via region-based and single-pass convolutional neural networks with comparison to conventional image analysis methodologies

As the complexity of microfluidic experiments and the associated image data volumes scale, traditional feature extraction approaches begin to struggle at both detection and analysis pipeline throughput. Deep-neural networks trained to detect certain objects are rapidly emerging as data gathering tools that can either match or outperform the analysis capabilities of the conventional methods used in microfluidic emulsion science. We demonstrate that two types of neural-networks, You Only Look Once (YOLOv3, YOLOv5) and Faster R-CNN, can be trained on a dataset which comprises of droplets generated across several microfluidic experiments and systems. The latitude of droplets used for training and validation, produce model weights which are easily transitive to emulsion systems at large, while completely circumventing any necessity of manual feature extraction. In flow cell experiments which comprised of greater than either 10,000 mono- or polydisperse droplets, the models show excellent or superior statistical symmetry to classical implementations of the Hough transform or widely utilized ImageJ plugins. In more complex chip architectures which simulate porous media, the produced image data typically requires heavy pre-processing to extrapolate valid data, where the models were able to handle raw input and produce size distributions with accuracy of ± 2μm for intermediate magnifications. This data harvesting fidelity is extended to foreign datasets not included in the training such as micrograph observation of various emulsified systems. Implementing these neural networks as the sole feature extraction tools in these microfluidic systems not only makes the data pipelining more efficient but opens the door for live detection and development of autonomous microfluidic experimental platforms due to inference times of greater than 100 frames per second.

N-gram MalGAN: Evading machine learning detection via feature n-gram

In recent years, many adversarial malware examples with different feature strategies, especially GAN and its variants, have been introduced to handle the security threats, e.g., evading the detection of machine learning detectors. However, these solutions still suffer from problems of complicated deployment or long running time. In this paper, we propose an n-gram MalGAN method to solve these problems. We borrow the idea of n-gram from the Natural Language Processing (NLP) area to expand feature sources for adversarial malware examples in MalGAN. Generally, the n-gram MalGAN obtains the feature vector directly from the hexadecimal bytecodes of the executable file. It can be implemented easily and conveniently with a simple program language (e.g., C++), with no need for any prior knowledge of the executable file or any professional feature extraction tools. These features are functionally independent and thus can be added to the non-functional area of the malicious program to maintain its original executability. In this way, the n-gram could make the adversarial attack easier and more convenient. Experimental results show that the evasion rate of the n-gram MalGAN is at least 88.58% to attack different machine learning algorithms under an appropriate group rate, growing to even 100% for the Random Forest algorithm.

Enhanced Radiomics for Prediction of Rupture Status in Cerebral Aneurysms

Radiomics is a powerful tool for automatic extraction of morphological features, but when applied to cerebral aneurysms, it is inferior to established descriptors in classifying rupture status. We sought a strategy to recover neck orientation and parent vessel caliber to enhance Radiomics performance and facilitate its adoption for aneurysm risk stratification. We analyzed 135 sidewall (32 ruptured) and 216 bifurcation (90 ruptured) aneurysms from three-dimensional rotational catheter angiography datasets. Clinical three-dimensional rotational catheter angiography defined in arbitrary orientation underwent affine transformations enabling aneurysm neck alignment to XY plane before analysis in PyRadiomics, facilitating automatic extraction of aneurysm height and width, previously not possible with random alignment. Additionally, parent vessel size was estimated from aneurysm location and incorporated into enhanced Radiomics (height, width, height/width, size ratio). Rupture status classification was compared across methodologies for 31 automatically computed conventional Radiomics, enhanced Radiomics, and established size/shape descriptors using univariate, multivariate, and area under the curve (AUC) statistics. Enhanced Radiomics-derived height/width and size ratio were significantly higher in both ruptured subsets. Using multivariate analysis in sidewall lesions, enhanced Radiomics (AUC= 0.85) matched established features (AUC= 0.86) and outperformed conventional Radiomics (AUC= 0.82); in bifurcation lesions, enhanced Radiomics (AUC= 0.78) outperformed both established features (AUC= 0.76) and conventional Radiomics (AUC= 0.74). Enhanced Radiomics incorporating neck orientation and parent vessel estimate is an efficient operator-independent methodology that offers superior rupture status classification for both sidewall and bifurcation aneurysms and should be considered a strong candidate for larger-scale multicenter and multimodality validation.

SEMI-AUTOMATED DELINEATION OF INFORMAL SETTLEMENT STRUCTURES FROM DRONE RGB IMAGERY USING OBJECT-BASED IMAGE ANALYSIS

Abstract. With the problem of informal settlements in the Philippines, mapping such areas is the first step towards improvement. Object-based image analysis (OBIA) has been a powerful tool for mapping and feature extraction, especially for high-resolution datasets. In this study, an informal settlement area in UP Diliman, Quezon City was chosen to be the subject site, where individual informal settlement structures (ISS) were delineated and estimated using OBIA. With the help of photogrammetry and image enhancement techniques, derivatives such as elevation model and orthophotos were produced for easier interpretation. An initial rule-set was developed to remove all non-ISS features from the base image–utilizing spectral values and thematic layers as main classifiers. This classification technique yielded a 94% accuracy for non-ISS class, and 92% for the possible ISS class. Another rule-set was then developed to delineate individual ISS based on the texture and elevation model of the area, which paved the way for the estimation of ISS count. To test the robustness of the methodology developed, the estimation results were compared to the manual count obtained through an online survey form, and the classification and delineation results were assessed through overall and individual quality checks. The estimation yielded a relative accuracy of 60%, which came from the delineation rate of 63%. On the other hand, delineation accuracy was calculated through area-based and number-based measures, yielding 58% and 95%, respectively. Issues such as noisy elevation models and physical limitations of the area and survey done affected the accuracy of the results.

Feature extraction for functional time series: Theory and application to NIR spectroscopy data

We propose a novel method to extract global and local features of functional time series. The global features concerning the dominant modes of variation over the entire function domain, and local features of function variations over particular short intervals within function domain, are both important in functional data analysis. Functional principal component analysis (FPCA), though a key feature extraction tool, only focus on capturing the dominant global features, neglecting highly localized features. We introduce a FPCA-BTW method that initially extracts global features of functional data via FPCA, and then extracts local features by block thresholding of wavelet (BTW) coefficients. Using Monte Carlo simulations, along with an empirical application on near-infrared spectroscopy data of wood panels, we illustrate that the proposed method outperforms competing methods including FPCA and sparse FPCA in the estimation functional processes. Moreover, extracted local features inheriting serial dependence of the original functional time series contribute to more accurate forecasts. Finally, we develop asymptotic properties of FPCA-BTW estimators, discovering the interaction between convergence rates of global and local features.

A Sample Preparation Workflow for Delayering a 45 nm Node Serial Peripheral Interface Module

Abstract Further development of SEM-based feature extraction tools for design validation and failure analysis is contingent on reliable sample preparation methods. This article describes how a delayering framework for 130 nm technology was adapted and used on a 45 nm SPI module consisting of 11 metal layers, 10 via layers, two layers of polysilicon, and an active silicon layer. It explains how different polishing and etching methods are used to expose each layer with sufficient contrast for SEM imaging and subsequent feature extraction. By combining polygon sets representing each layer, the full design of the device was reconstructed as shown in one of the images.

A Comparative Study of Protein Sequences Classification-Based Machine Learning Methods for COVID-19 Virus against HIV-1

ABSTRACT The effective spread of COVID-19 cases in several countries produces more protein sequences that are released in genomic public sources. It provides some awareness and indications for virus classification of COVID-19 and HIV-1 that are essential for drug discovery of COVID-19. This paper reveals the importance of machine learning algorithms to handle the recognition of two different viruses. Therefore, 18,476 protein sequences for both COVID-19 and HIV-1 and 9238 for each virus are applied to the proposed model based on feature extraction, data labeling, and six classifiers. Amino acid classification according to their dipoles and volumes is employed as a feature extraction tool based on the creation of eight features from twenty amino acids by using the conjoint triad (CT) method. The data labeling is employed as a coding tool by binary numbers refereeing zero for COVID-19 and one for HIV-1. The random forest (RF) model achieved the highest classification accuracy of 99.89% for eight features and 97.80% for two features. The experimental results significantly confirmed that eight features required more computational time than two features, but the accuracy rate was nearly similar in the two cases. This classification strategy of COVID-19 and HIV-1 will prompt the prediction of protein sequences of the new virus.

Prognosis and Prediction of Breast Cancer Using Machine Learning and Ensemble-Based Training Model

Abstract There has been an increase in occurrence of human diseases all over the world. Among those, Breast Cancer has increased with an alarming rate in the past decade and this trend of increase would continue to grow. Now, there is a need for efficient text analytics and feature extraction tools to assist classifying, sharing and retrieving the information on human diseases in general and Breast Cancer in particular. In light of above, the present study has been undertaken with the objective to provide a comparative analysis of different classifiers on Breast Cancer dataset, and to propose a new ensemble training method of Machine Learning Classification. Here, machine learning models (such as K-Nearest Neighbour, Logistic Regression, Decision Tree, Random Forest, Gradient Boost, Support Vector Machine) and deep learning classifiers (such as Multi-Layer Feed Forward Neural Network, Recurrent Neural Network and Long Short Term Memory) have been applied on Breast Cancer dataset. An Ensemble Learning model for Prediction is proposed to classify the results among different classifiers. Finally, the Voting Ensemble is implemented to find out the optimal classifier for prediction of Breast Cancer. The results have been computed using the evaluation parameters such as Accuracy, Precision, Recall and Specificity. The confusion matrix drawn on the basis of evaluation parameters provides more emphasis on predicted and actual instances. Performance Evaluation for various machine learning models is computed. Results of this investigation concludes that Voting Ensemble outperforms other machine learning models. The prediction using Voting Ensemble resulted in an accuracy rate of 97.9 per cent, precision of 96.77 per cent and recall of 100 per cent.

Transliteration of Hiragana and Katakana Handwritten Characters Using CNN-SVM

Hiragana and katakana handwritten characters are often used when writing words in Japanese. Japanese itself is often used by native Japanese as well as people learning Japanese around the world. Hiragana and katakana characters themselves are difficult to learn because many characters are similar to one another. In this study, hiragana and basic katakana, dakuten, handakuten, and youon were used, which were taken from the respondents using a questionnaire. This study used the CNN method which will be compared with a combination of the CNN and SVM methods which have been designed to identify each character that has been prepared. Preprocessing of character images uses the methods of image resizing, grayscaling, binarization, dilation, and erosion. The preprocessed results will be input for CNN as a feature extraction tool and SVM as a tool for character recognition. The results of this study obtained accuracy with the following parameters: 69×69 image size, 3 patience values, val_loss monitor callbacks, Nadam optimization function, 0.001 learning rate value, 30 epochs value, and SVM RBF kernel. If using a system that only uses the CNN network, the accuracy is 87.82%. The results obtained when using a combination of CNN and SVM were 88.21%.

Introducing the Discrete Path Transform (DPT) and its applications in signal analysis, artefact removal, and spoken word recognition

This article introduces the Discrete Path Transform (DPT). Designed to serve as a new tool for handcraft feature extraction (FE), it improves the elementary analysis provided by signal energy (E) and enhances the humble spectral investigation granted by zero-crossing rates (ZCRs). C/C++ source-codes to realize both the DPT direct and inverse (IDPT) forms are presented together with a few hypothetical numerical examples and an application involving general signal analysis, artefact removal from biomedical signals, and spoken word recognition (SWR), thus demonstrating how useful and effective the proposed transform is. Brief comparisons with Teager Energy Operator (TEO) and a list of important references were also included.