Feature Set For Classification Research Articles

An Improved Feature Set for Hyperspectral Image Classification: Harmonic Analysis Optimized by Multiscale Guided Filter

Effective features derived from an original hyperspectral image (HSI) are quite important to improve the classification performance. An improved feature set, namely HGFM, is constructed by integrating harmonic analysis (HA) optimized by a multiscale guided filter (GF) with morphological operation for HSI classification. To establish HGFM, HA is first adopted to convert the HSI from spectral space to the frequency domain represented by amplitude, phase, and residual. With the first component of minimum noise fraction obtained from the original HSI as the guidance image, the harmonic components are then processed by the multiscale GF. Finally, the obtained results are then operated via morphological opening by reconstruction and closing by reconstruction to generate an improved feature set for classification. The HGFM features are input to an ensemble learning (EL) based on classification framework, in which EL plays an auxiliary role to enhance the classification stability and reliability. Three commonly used HSIs are used for experiments, and different feature sets are evaluated by comparing EL and rotation forest, support vector machine optimized by particle swarm optimization, random forest, and others. Compared with benchmark feature sets, the proposed HGFM feature set can better depict the details of objects easily, and the experimental results confirm the effectiveness in terms of classification accuracy and generalization ability.

Learning Markov Blankets From Multiple Interventional Data Sets.

Learning Markov blankets (MBs) plays an important role in many machine learning tasks, such as causal Bayesian network structure learning, feature selection, and domain adaptation. Since variables included in the MB of a target variable of interest have causal relationships with the target, the MB can serve as the basis of learning the global structure of a causal Bayesian network or as a reliable and robust feature set for classification, both within the same domain or across domains. In this article, we study the problem of learning the MB of a target variable from multiple interventional data sets. Data sets attained from interventional experiments contain richer causal information than passively observed data (observational data) for MB discovery. However, almost all existing MB discovery methods are designed for learning MBs from a single observational data set. To learn MBs from multiple interventional data sets, we face two challenges: 1) unknown intervention variables and 2) nonidentical data distributions. To address these challenges, we theoretically analyze: 1) under what conditions we can find the correct MB of a target variable and 2) under what conditions we can identify the causes of the target variable via discovering its MB. Based on the theoretical analysis, we propose a new algorithm for learning MBs from multiple interventional data sets, and we present the conditions/assumptions that assure the correctness of the algorithm. To the best of our knowledge, this article is the first to present the theoretical analyses about the conditions for MB discovery in multiple interventional data sets and the algorithm to find the MBs in relation to the conditions. Using benchmark Bayesian networks and real-world data sets, the experiments have validated the effectiveness and efficiency of the proposed algorithm in this article.

Comparing Multiclass, Binary, and Hierarchical Machine Learning Classification schemes for variable stars

ABSTRACTUpcoming synoptic surveys are set to generate an unprecedented amount of data. This requires an automatic framework that can quickly and efficiently provide classification labels for several new object classification challenges. Using data describing 11 types of variable stars from the Catalina Real-Time Transient Survey (CRTS), we illustrate how to capture the most important information from computed features and describe detailed methods of how to robustly use information theory for feature selection and evaluation. We apply three machine learning algorithms and demonstrate how to optimize these classifiers via cross-validation techniques. For the CRTS data set, we find that the random forest classifier performs best in terms of balanced accuracy and geometric means. We demonstrate substantially improved classification results by converting the multiclass problem into a binary classification task, achieving a balanced-accuracy rate of ∼99 per cent for the classification of δ Scuti and anomalous Cepheids. Additionally, we describe how classification performance can be improved via converting a ‘flat multiclass’ problem into a hierarchical taxonomy. We develop a new hierarchical structure and propose a new set of classification features, enabling the accurate identification of subtypes of Cepheids, RR Lyrae, and eclipsing binary stars in CRTS data.

EEG Signals Classification Using Machine Learning for The Identification and Diagnosis of Schizophrenia.

This paper presents the design of a machine learning-based classifier for the differentiation between Schizophrenia patients and healthy controls using features extracted from electroencephalograph(EEG) signals based on event related potential(ERP). A number of features are extracted from an online EEG dataset with 81 subjects, including 32 healthy controls and 49 Schizophrenia patients. The EEG signals are preprocessed and since the dataset is relatively small, the random forest machine learning algorithm is chosen to be applied on different combinations of feature sets for classification. It is found that the classification accuracy can be improved by adding certain features extracted from EEG signals.

Mapping Tea Plantations from Multi-seasonal Landsat-8 OLI Imageries Using a Random Forest Classifier

It is very challenging to extract tea plantations from medium-resolution satellite imageries. This paper presents a new methodological framework based on the random forest classifier for extracting tea plantations from Landsat-8 OLI imageries. Analysis is facilitated by a dataset of three Landsat-8 OLI images (spring, autumn and winter) in 2014 covering the Anji County, one of the major tea production regions in China. More specifically, in order to determine the relative importance of spectra, texture, vegetation index and seasonality on classification accuracy, we design a series of classification feature sets, including: initial feature sets for different seasons, feature selection feature sets for different seasons and multi-seasonal feature sets. The results show that the multi-seasonal feature selection set has the best feature set performance (PA = 0.88; OA = 0.92; Kappa = 0.90). Our study demonstrates that the random forest classifier is reliable and practical for extracting tea plantations from medium-resolution images. Highlights of this study are: mapping tea plantations, an important cash crop to local agriculture, in a fragmented landscape; integrating textual, vegetational, and seasonal features, which were helpful for improving tea plantation mapping accuracy in combination; taking advantage of the feature selection function of random forest, supporting high-dimensional data classification, which leads to a higher classification accuracy.

Finger movements recognition using minimally redundant features of wavelet denoised EMG

Developing prosthetic hands with high functionality and ease of use is the focus of current research in the area of Electromyogram (EMG) based prosthesis control. Although individuals with upper limb loss can perform grasping operations with currently available prosthetic hands, more intuitive control of finger movements is required to replicate the complex motor functions of human hands. A significant challenge is to classify the finger movements with higher recognition rates using a smaller number of EMG channels. This paper reports a novel criterion for selection of minimally redundant EMG feature set for classification of 10-class finger movements using two-channel EMG. The feature set is selected from wavelet denoised EMG at four decomposition levels using minimum redundancy in terms of mutual information. A set of five features: root mean square, simple square integral, slope sign change, peak frequency and power spectral ratio have been selected from 31 time and frequency domain features. Using the current state of the art classification technique based on support vector machine, we achieved 10-fold cross-validation recognition rate of 96.5 ± 0.13%. The experimental study shows that the feature set with minimum redundancy in terms of mutual information ensures the highest recognition rate with reduced computational cost.

Classifying Periodic Astrophysical Phenomena from non-survey optimized variable-cadence observational data

Modern time-domain astronomy is capable of collecting a staggeringly large amount of data on millions of objects in real time. Therefore, the production of methods and systems for the automated classification of time-domain astronomical objects is of great importance. The Liverpool Telescope has a number of wide-field image gathering instruments mounted upon its structure, the Small Telescopes Installed at the Liverpool Telescope. These instruments have been in operation since March 2009 gathering data of large areas of sky around the current field of view of the main telescope generating a large dataset containing millions of light sources. The instruments are inexpensive to run as they do not require a separate telescope to operate but this style of surveying the sky introduces structured artifacts into our data due to the variable cadence at which sky fields are resampled. These artifacts can make light sources appear variable and must be addressed in any processing method.The data from large sky surveys can lead to the discovery of interesting new variable objects. Efficient software and analysis tools are required to rapidly determine which potentially variable objects are worthy of further telescope time. Machine learning offers a solution to the quick detection of variability by characterising the detected signals relative to previously seen exemplars. In this paper, we introduce a processing system designed for use with the Liverpool Telescope identifying potentially interesting objects through the application of a novel representation learning approach to data collected automatically from the wide-field instruments. Our method automatically produces a set of classification features by applying Principal Component Analysis on set of variable light curves using a piecewise polynomial fitted via a genetic algorithm applied to the epoch-folded data. The epoch-folding requires the selection of a candidate period for variable light curves identified using a genetic algorithm period estimation method specifically developed for this dataset. A Random Forest classifier is then used to classify the learned features to determine if a light curve is generated by an object of interest. This system allows for the telescope to automatically identify new targets through passive observations which do not affect day-to-day operations as the unique artifacts resulting from such a survey method are incorporated into the methods.We demonstrate the power of this feature extraction method compared to feature engineering performed by previous studies by training classification models on 859 light curves of 12 known variable star classes from our dataset. We show that our new features produce a model with a superior mean cross-validation F1 score of 0.4729 with a standard deviation of 0.0931 compared with the engineered features at 0.3902 with a standard deviation of 0.0619. We show that the features extracted from the representation learning are given relatively high importance in the final classification model. Additionally, we compare engineered features computed on the interpolated polynomial fits and show that they produce more reliable distributions than those fit to the raw light curve when the period estimation is correct.

Identification of Self-Admitted Technical Debt Using Enhanced Feature Selection Based on Word Embedding

Self-admitted technical debt (SATD) is annotated in source code comments by developers and has been recognized as a great source of discovering flawed software. To reduce manual effort, some recent studies have focused on automated detection of SATD using text classification methods. To train their classifier, these methods need labeled samples, which also require a lot of effort to obtain. We developed a new SATD identification method, which takes advantage of a large corpus of unlabeled code comments, extracted from open source projects, to train a word embedding model. After applying feature selection, the pre-trained word embedding is used for discovering semantically similar features in source code comments to enhance the original feature set. By using such enhanced feature set for classification, our goal was to improve the identification of SATD when compared to existing methods. The proposed feature enhancement method was used with the three most common feature selection methods (CHI, IG, and MI), and three well-known text classification algorithms (NB, SVM, and ME) and was tested on ten open source projects. The experimental results show a significant improvement in SATD identification over the compared methods. With an achieved 82% of correct predictions of SATD, the proposed method seems to be a good candidate to be adopted in practice.

Genetik Algoritmalar Kullanarak EKG Vuru Sınıflandırması için Öznitelik Seçimi

In this study, genetic algorithm method was used to select the most suitable set of features for classification of arrhythmia types of heart beats. Normal, right branch block, left branch block and pace rhythm samples of electrocardiography (ECG) signals which obtained from the MIT-BIH cardiac arrhythmia database were used in the classification. Mean, standard deviation, energy and entropy of discrete wavelet transform (DWT) coefficients were proposed as the features for the classification. By using the proposed DWT method, 16 features which have high classification accuracy were obtained among the 208 feature sets constructed from 13 different wavelet types by applying the genetic algorithm method. It was observed that the features that increase accuracy can be detected by the genetic algorithm and the feature set obtained from the coefficients of the different types of wavelets selected at different levels show higher performance than the coefficients obtained from the standard individual wavelet in the ECG arrhythmia classification.

Multi-co-training for document classification using various document representations: TF–IDF, LDA, and Doc2Vec

The purpose of document classification is to assign the most appropriate label to a specified document. The main challenges in document classification are insufficient label information and unstructured sparse format. A semi-supervised learning (SSL) approach could be an effective solution to the former problem, whereas the consideration of multiple document representation schemes can resolve the latter problem. Co-training is a popular SSL method that attempts to exploit various perspectives in terms of feature subsets for the same example. In this paper, we propose multi-co-training (MCT) for improving the performance of document classification. In order to increase the variety of feature sets for classification, we transform a document using three document representation methods: term frequency–inverse document frequency (TF–IDF) based on the bag-of-words scheme, topic distribution based on latent Dirichlet allocation (LDA), and neural-network-based document embedding known as document to vector (Doc2Vec). The experimental results demonstrate that the proposed MCT is robust to parameter changes and outperforms benchmark methods under various conditions.

Machine learning models to predict the progression from early to late stages of papillary renal cell carcinoma

Papillary Renal Cell Carcinoma (PRCC) is a heterogeneous disease with variations in disease progression and clinical outcomes. The advent of next generation sequencing techniques (NGS) has generated data from patients that can be analysed to develop a predictive model. In this study, we have adopted a machine learning approach to identify biomarkers and build classifiers to discriminate between early and late stages of PRCC from gene expression profiles. A machine learning pipeline incorporating different feature selection algorithms and classification models is developed to analyse RNA sequencing dataset (RNASeq). Further, to get a reliable feature set, we extracted features from different partitions of the training dataset and aggregated them into feature sets for classification. We evaluated the performance of different algorithms on the basis of 10-fold cross validation and independent test dataset. 10-fold cross validation was also performed on a microarray dataset of PRCC. A random forest based feature selection (varSelRF) yielded minimum number of features (104) and a best performance with area under Precision Recall curve (PR-AUC) of 0.804, MCC (Matthews Correlation Coefficient) of 0.711 and accuracy of 88% with Shrunken Centroid classifier on a test dataset. We identified 80 genes that are consistently altered between stages by different feature selection algorithms. The extracted features are related to cellular components - centromere, kinetochore and spindle, and biological process mitotic cell cycle. These observations reveal potential mechanisms for an increase in chromosome instability in the late stage of PRCC. Our study demonstrates that the gene expression profiles can be used to classify stages of PRCC.

Cost and managerial accounting of joint activities related to the use of fixed assets

The research paper reveals the essence of the concepts of joint activity, joint operation, and joint venture. A set of key features for classification of joint activities is identified and their impact on accounting of joint activities is assessed. The article also reviews the essential elements of accounting of joint activities in the light of International Financial Reporting Standards (IFRS), and characterizes the process of recording accounting entries related to basic operations, which depend on organizational forms of joint activities (a joint venture or a joint operation, with or without a separate entity). The paper provides a detailed description of three options for accounting of joint activities classified as joint operations, namely: joint operations without a separate entity; joint operations with a separate entity but without legal personality; a legal unit. Besides, a number of particular characteristics of measuring financial results from selling and purchasing assets within joint operations are identified. It is pointed out that one of the ways of effective use of fixed assets is promoting the implementation of managerial ac- counting of joint activities and internal reporting procedures of the results achieved. It is suggested that domestic enterprises of oil and gas industry should expand the practice of joint activities in order to effectively use fixed assets for oil and gas extraction and transportation. Before conducting joint activities, it is recommended that oil and gas industry enterprises compile initial calculations of their profitability at the level of managerial accounting.In the study, the following general and specific scientific methods of obtaining knowledge on economic phenomena are used: generalization, grouping and comparison, analysis, synthesis, induction and deduction, etc.

Improved lung nodule diagnosis accuracy using lung CT images with uncertain class

Background and objective: Among all malignant tumors, lung cancer ranks in the top in mortality rate. Pulmonary nodule is the early manifestation of lung cancer, and plays an important role in its discovery, diagnosis and treatment. The technology of medical imaging has encountered a rapid development in recent years, thus the amount of pulmonary nodules can be discovered are on the raise, which means even tiny or minor changes in lung can be recorded by the CT images. This paper proposes a pulmonary nodule computer aided diagnosis (CAD) based on semi-supervised extreme learning machine(SS-ELM).Methods: First, the feature model based on the pulmonary nodules regions of lung CT images is established. After that, the same feature data sets have been put into ELM, support vector machine (SVM) methods, probabilistic neural network (PNN) and multilayer perceptron (MLP) so as to compare the performance of the methods. ELM turned out to have better performance in training time and testing accuracy compared with SVM, PNN and MLP. Then, we propose a pulmonary nodules computer aided diagnosis algorithm based on semi-supervised ELM (SS-ELM), which enables both certain class feature sets with labels and unlabeled feature sets to be input for training and computer aided diagnosing. This algorithm has provided a solution for the using of uncertain class data and improve the testing accuracy of benign and malignant diagnosis.Results: 1018 sets of thoracic CT images from the Lung Database Consortium and Image Database Resource Initiative (LIDC-IDRI) have been used in experiment in order to test the effectiveness of the algorithm. Compared with ELM, the pulmonary nodules CAD based on SS-ELM has better testing accuracy performance.Conclusions: We have proposed a pulmonary nodule CAD system based on SS-ELM, which achieving better generalization performance at faster learning speed and higher testing accuracy than ELM, SVM, PNN and MLP. The SS-ELM based pulmonary nodules CAD has been proposed to solve the problem of uncertain class data using.

Classifying airborne LiDAR point clouds via deep features learned by a multi-scale convolutional neural network

ABSTRACTPoint cloud classification plays a critical role in many applications of airborne light detection and ranging (LiDAR) data. In this paper, we present a deep feature-based method for accurately classifying multiple ground objects from airborne LiDAR point clouds. With several selected attributes of LiDAR point clouds, our method first creates a group of multi-scale contextual images for each point in the data using interpolation. Taking the contextual images as inputs, a multi-scale convolutional neural network (MCNN) is then designed and trained to learn the deep features of LiDAR points across various scales. A softmax regression classifier (SRC) is finally employed to generate classification results of the data with a combination of the deep features learned from various scales. Compared with most of traditional classification methods, which often require users to manually define a group of complex discriminant rules or extract a set of classification features, the proposed method has the ability to automatically learn the deep features and generate more accurate classification results. The performance of our method is evaluated qualitatively and quantitatively using the International Society for Photogrammetry and Remote Sensing benchmark dataset, and the experimental results indicate that our method can effectively distinguish eight types of ground objects, including low vegetation, impervious surface, car, fence/hedge, roof, facade, shrub and tree, and achieves a higher accuracy than other existing methods.

Classification of patients with tumor using MR FLAIR images

Magnetic Resonance Imaging (MRI) is a fast growing imaging tool for neurodiagnosis. The radiologists time is at a premium due to increase in patient studies each having a large data set. This can be aided by classification using machine learning techniques. This paper evaluates its utility for accurate and rapid diagnosis of cerebral tumors. Two hundred subjects were classified into normal and abnormal using volumetric Fluid Attenuated Inversion Recovery (FLAIR) acquisition. The images are normalized to obtain 12 useful slices to be considered as the patient feature set for classification. Discrete Wavelet Transform (DWT) is used for feature extraction and Principal Component Analysis (PCA) is used for feature selection. Various classifiers like Support Vector Machine (SVM), k-Nearest Neighbor (k-NN), CART (Classification and Regression Tree) and Random forest are tested. Applying K-fold cross validation in each train-test ratios, we obtained ceiling level classification accuracy with coherent sensitivity and specificity using only linear SVM (negating the use of PCA). An accuracy of 88% is obtained with a sensitivity of 84% and specificity of 92% with 62.28 s computation time. The algorithm is robust to be tested in clinical settings.

P.1.a.003 - PDE7B, NMBR and EPM2A variants and schizophrenia: a case-control and pharmacogenetics study

Using a large set of ultrasound features does not necessarily ensure improved quantitative classification of breast tumors; rather, it often degrades the performance of a classifier. In this paper, we propose an effective feature reduction approach in the transform domain for improved multi-class classification of breast tumors. Feature transformation methods, such as empirical mode decomposition (EMD) and discrete wavelet transform (DWT), followed by a filter- or wrapper-based subset selection scheme are used to extract a set of non-redundant and more potential transform domain features through decorrelation of an optimally ordered sequence of N ultrasonic bi-modal (i.e., quantitative ultrasound and elastography) features. The proposed transform domain bi-modal reduced feature set with different conventional classifiers will classify 201 breast tumors into benign-malignant as well as BI-RADS ⩽ 3, 4, and 5 categories. For the latter case, an inadmissible error probability is defined for the subset selection using a wrapper/filter. The classifiers use train truth from histopathology/cytology for binary (i.e., benign-malignant) separation of tumors and then bi-modal BI-RADS scores from the radiologists for separating malignant tumors into BI-RADS category 4 and 5. A comparative performance analysis of several widely used conventional classifiers is also presented to assess their efficacy for the proposed transform domain reduced feature set for classification of breast tumors. The results show that our transform domain bimodal reduced feature set achieves improvement of 5.35%, 3.45%, and 3.98%, respectively, in sensitivity, specificity, and accuracy as compared to that of the original domain optimal feature set for benign-malignant classification of breast tumors. In quantitative classification of breast tumors into BI-RADS categories ⩽ 3, 4, and 5, the proposed transform domain reduced feature set attains improvement of 3.49%, 9.07%, and 3.06%, respectively, in likelihood of malignancy and 4.48% in inadmissible error probability compared to that of the original domain optimal subset. In summary, the construction of a transform domain reduced feature set by extracting complementary information from a large set of available bi-modal features and use of qualitative bi-modal BI-RADS can contribute to improved quantitative classification of breast tumors and thereby help reduce the number of unnecessary biopsies, securing a nearly minimum chance of a life-endangering diagnosis.

Review of Methods for Synthesis of the Model of Class of Spiroid Gearboxes for Intelligent CAD Systems. Part 1. Graph Method

Создание интеллектуальных систем автоматизации конструирования требует решения задачи накопления и обобщения конструкторского опыта. Разработка модели конструкторской деятельности сталкивается с отсутствием сквозного алгоритма, позволяющего синтезировать конструкцию, имея в качестве исходных данных потребительские качества изделия. Именно этим объясняется функциональный набор современных САПР, которые способны автоматизировать процессы выполнения графических работ, строить трехмерные модели, выполнять отдельные инженерные расчеты, готовить конструкторскую и технологическую документацию, однако не позволяют выполнить структурный синтез будущего изделия. Решающим фактором в реализации процесса структурного синтеза становятся опыт конструктора и его интуиция. Автор статьи, разрабатывая методологию автоматизации процесса структурного синтеза, ставит вопрос о накоплении «конструкторского опыта» для его дальнейшего использования в качестве информационного обеспечения системы автоматизированного конструирования, способной максимально автоматизировать процесс структурного синтеза будущей конструкции. Обобщение опыта рассматривается автором как формирование модели класса объектов, позволяющей зафиксировать максимальное количество особенностей конструктивных исполнений изделия и его компонентов. Еще одна задача, которая должна быть решена при создании модели класса объектов, - это обеспечение возможности ее автоматизированной реализации. В качестве модели класса объектов автор предлагает использовать классификатор, представляющий собой набор классификационных признаков (вопросов) и множество значений признаков (альтернативных ответов на каждый из поставленных вопросов). В качестве объекта конструирования рассмотрен спироидный редуктор, представляющий собой объект средней степени сложности. Графовое представление конструкции объекта, позволяющее продемонстрировать все конструктивные единицы, обладающие определенными, интересующими конструктора свойствами, является наглядным и удобным для математического, а следовательно, алгоритмического описания. В статье обоснована необходимость и возможность создания классификатора в качестве обобщенной модели класса спироидных редукторов, построенной графовым методом, состоящим в обобщении графовых моделей отдельных конструктивных решений. Автором рассмотрен графовый метод Половинкина, обоснована необходимость его модификации, предложен графовый метод, обеспечивающий высокую степень автоматизации процесса обобщения конструкторского опыта и создания классификатора.

Automatic recognition of electric loads analyzing the characteristic parameters of the consumed electric power through a Non-Intrusive Monitoring methodology

Non-Intrusive Load Monitoring (NILM) consists in measuring the electricity consumption using a power consumption data acquisition system, typically placed in the main supply of the building. Relying on a single point of measure it is also called one-sensor metering in contrast to the common metering hardware that can be embedded in each appliance (electronic metering ore-metering) and in differentiation with the common utility smart meters. NILM is the process in which you are able to disaggregate a set of energy readings over a period of time to determine exactly what appliances have used the power and how much power each appliance has used during that time period. In this work a supervised classification method was employed for offline appliances classification, based on low frequency power consumption. The classification feature set consists of the true power, reactive power, and the step changing of the true power. Multiple classifiers were tested and evaluated, such as Decision Tree, Nearest Neighbor, Discriminant Analysis, and the multilayer Feed-forward Neural Network classifier. The methods were tested on the ACS-F2 appliance consumption signatures database.

Unsupervised Data Driven Feature Extraction by Means of Mutual Information Maximization

In Earth observations technical literature, several methods have been proposed and implemented to efficiently extract a proper set of features for classification and segmentation purposes. However, these architectures show drawbacks when the considered datasets are characterized by complex interactions among the samples, especially when they rely on strong assumptions on noise and label domains. In this paper, a new unsupervised approach for feature extraction, based on data driven discovery, is introduced for accurate classification of remotely sensed data. Specifically, the proposed architecture exploits mutual information maximization in order to retrieve the most relevant features with respect to information measures. Experimental results on real datasets show that the proposed approach represents a valid framework for feature extraction from remote sensing images.

An Approach to Text Classification based on Age Groups of Addressees

The article deals with a new approach to text classification considering the existence of different types of classification features (binary, nominal, ordinal and interval). The specialty of the approach is a phased classification process, which makes it possible to not cause different types of classification features to a single range. The author describes a computational experiment using texts included in Russian National Corpus and suggests the set of classification features for Russian text classification based on the age of theirs supposed readers. Text documents included in the sample are divided into two categories – for adults and for children, — according to the views of experts.