Characteristics Of Different Classes Research Articles

Learn Generalization Feature via Convolutional Neural Network: A Fault Diagnosis Scheme Toward Unseen Operating Conditions

In recent years, Convolutional neural networks (CNNs) have achieved start-of-art performance in the fault diagnosis field. If there is no available information on the unseen operating conditions, the model trained on the seen operating condition cannot perform well. One of the feasible strategies is to enhance the generalization ability of the network on various seen operating conditions. We introduce the center loss to the traditional CNN and build an end-to-end fault diagnosis framework (called CNN-C). By minimizing the intra-class variations, center loss cluster the learned features across various seen operating conditions. With the joint supervision of the center loss and the softmax loss, the learned features of the same class could minimize the domain difference across various seen operating conditions while the features of different classes are separable. The generalization ability of network is improved on unseen operating conditions. Compared with the shallow methods and traditional CNN, the proposed method is promising to deal with the fault diagnosis tasks of the bearing and gearbox.

Butterfly identification using gray level co-occurrence matrix (glcm) extraction feature and k-nearest neighbor (knn) classification

Gita Persada Butterfly Park is the only breeding of engineered in situ butterflies in Indonesia. It is located in Lampung and has approximately 211 species of breeding butterflies. Each species of Butterflies has a different texture on its wings. The Limited ability of the human eye to distinguishing typical textures on butterfly species is the reason for conducting a research on butterfly identification based on pattern recognition. The dataset consists of 600 images of butterfly’s upper wing from six species: Centhosia penthesilea, Papilio memnon, Papilio nephelus, Pachliopta aristolochiae, Papilio peranthus and Troides helena. The pre-processing stage is conducted using scaling, segmentation and grayscale methods. The GLCM method is used to recognize the characteristics of butterfly images using pixel distance and Angular direction 0o, 45o, 90o and 135o. The features used is angular second moment, contrast, homogeneity and correlation. KNN classification method in this study uses k values1, 3, 5, 7, 9, 11, 13, 15, 17, 19, 21 and 23 based on the Rule of Thumb. The result of this study indicate that Centhosia penthesilea and Papilio nephelus classes can be classified properly compared to the other 4 classes and require a classification time of 2 seconds at each angular orientation. The highest accuracy is 91.1% with a value of in the angle of 90o and error rate8.9%. Classification error occured because the value of the test data features is more dominant with the value of the training image features in different classes than the supposed class. Another reason is because of imperfect test data.

Image Feature Selection using Ant Colony Optimization

With the enhancement in imaging technology, huge amount of information is being created whose classification is a challenging task among researchers. The classification performance is solely dependent on the quality of the features defining the respective images. Hence, feature extraction from images and feature selection from this huge data is required. Image Feature Selection is a vital step which can significantly affect the performance of image classification system. This paper proposes an efficient combination of image feature extraction technique and feature selection strategy using Ant Colony Optimization (ACO). ACO utilizes overall subsets performance and local feature importance to fetch problem domain finding prime solutions. For ACO based feature selection in images most of the researchers use priori information of features. However, in this work, the features are extracted separately through the detailed analytical process which carries the best suited features for different classes of images. The performance in terms of classification accuracy has been enhanced through the optimization of features in the dataset. For experimentation and evaluating the performance of proposed work, Corel and Caltech datasets are used. Satisfactory results have been obtained with the proposed approach.

CBMT-07. BIOMARKERS OF AGGRESSIVENESS IN IDH MUTANT GLIOMA

Abstract BACKGROUND Gliomas with isocitrate dehydrogenase (IDH) mutations in adults evolve from lower-grade gliomas to secondary glioblastomas (GBM), a fatal disease with fast progression. IDH mutation occurs early in tumorigenesis, and persistently contribute to the reprograming of glucose, lipid and amino acid metabolism. This offer a plethora of potential biomarkers of progression. However, because it is extremely difficult to detect the distribution and transfer of metabolites changing in every moment in a single cell, the involvement of metabolites produced by mutant IDH in malignant progression remains understudied. MATERIALS AND METHODS Raman imaging spectroscopy, which can image chemical bonds and concentration of molecules at submicron spatial resolution, enables detection of spatiotemporal changes of metabolomes in live cells. We developed the software called Biomolecular Component Analysis (BCAbox) to deconvolute the recorded raw Raman spectra, leading to detection of unique spectral features of different classes of biomolecules. RESULTS AND CONCLUSIONS We applied Raman imaging spectroscopy to GBM cell lines that were transfected with IDH1 mutant gene. Our results indicated that lipid metabolism has a unique profile in IDH1 mutant gliomas. Subsequent mass spectrometry analysis of extracted organelle revealed the exact classes of lipids altered in the IDH mutant glioma and suggested biomarkers unique to IDH1 mutant. We will report our validation studies of the biomarkers in patient-derived IDH mutant glioma cell lines and patients derived-orthotopic xenograft mouse models with different degrees of aggressiveness and in matched primary versus recurrent gliomas. The results of the present study may provide novel insights into the discovery of metabolic biomarkers for the malignant progression in IDH mutant gliomas.

Higher order statistics for modulation and STBC recognition in MIMO systems

Identification of modulation and space-time block code (STBC) is an important task of receivers in applications such as military, civilian, and commercial communications. Here, we consider multiple-input multiple-output (MIMO) systems. We propose two methods for STBC identification when the modulation is known. We also introduce a method for joint identification of code and modulation. Additionally, we present an enhanced zero-forcing (ZF) equaliser to improve the separation between the features of different classes. Higher order cumulants are used as the statistical features. In the first method of STBC identification, after the proposed equalisation, received data samples are segmented, and then using the mean and Frobenius norm of the covariance matrix of extracted cumulants as threshold values, the STBCs are detected. This method requires the knowledge of the threshold values for each modulation type and noise power. In the second method, the knowledge of noise power and the type of modulation are not needed. After the proposed ZF equalisation, feature vector based on the cumulants is calculated, and then support vector machine (SVM) classifier is used to classify different STBCs. In the third method, joint detection of STBC and modulation type is performed using the proposed ZF equaliser and mapping the received samples. This method deals with the theoretical values calculated from the cumulants of four STBCs and four modulations, where there is no need to know the noise power. The results indicate that the proposed methods perform well even at low signal-to-noise ratios (SNRs).

Flow unit characteristics of fan delta front deposits and its influence on reservoir development - Taking yulou oil bearing sets in some experimental area in west depression in Liaohe Basin in China as an example

Taking some steam flooding experimental area in the western depression of the Liaohe Basin in China as an example, it quantitatively evaluates the flow units of the sedimentary facies in fan delta front and analyzes the influences of the flow units on the development of the oilfields. Comprehensive analysis of multifarious data such as geology, core analysis and well logs indicates that porosity, permeability and FZI are key parameters for classification and evaluation of the reservoir flow units. Then using SPSS cluster analysis software platform, cluster analysis are conducted on reservoir physical properties of cored wells, and based on which discriminant analysis are performed to obtain the discriminant formula. The discriminant analysis of reservoir physical properties of the wells without core data is based on fine interpretation of well logging data. Combined with the results of cored wells, the classification and evaluation of the flow units has been completed in the target layer of ​​the whole study area. Finally the flow units are divided into four classes, namely, Class A, Class B, Class C and Class D. Classes A and B flow units represent high-quality reservoirs. The flow units are mainly controlled by sedimentation. Classes A and B are mainly sedimentary microfacies of underwater distributary channel and distributary mouth sand bar. The paper proposes five principles of judging the rationality of flow units, describes the development characteristics of different classes of flow units on section and plane, analyzes the influence of flow units on oilfield development and suggests countermeasures. Specially, in the process of steam flooding, target injection layers should be optimized and steam injection pressure should be effectively controlled to prevent steam channeling; when planning infill wells, injection and production wells should be deployed in the same class of flow unit to ensure good connectivity; the development of remaining oil should focus on Classes A and B flow units, and meanwhile pay more attention to the fine interpretation results of well logging data.

Fusing High-Spatial-Resolution Remotely Sensed Imagery and OpenStreetMap Data for Land Cover Classification Over Urban Areas

Land cover classification of urban areas is critical for understanding the urban environment. High-resolution remotely sensed imagery provides abundant, detailed spatial information for urban classification. In the meantime, OpenStreetMap (OSM) data, as typical crowd-sourced geographical information, have been an emerging data source for obtaining urban information. In this context, a land cover classification method that fuses high-resolution remotely sensed imagery and OSM data is proposed. Training samples were generated by integrating the OSM data and multiple information indexes. OSM data, which contain class attributes and location information of urban objects, served as the labels of initial training samples. Multiple information indexes that reflect spectral and spatial characteristics of different classes were utilized to improve the training set. Morphological attribute profiles were used because the structural and contextual information of images was effective in distinguishing the classes with similar spectral characteristics. Moreover, a road superimposition strategy that considers road hierarchy was developed because OSM data provide road information with high completeness in the urban area. Experiments were conducted on the data captured over Wuhan city, and three state-of-the-art approaches were adopted for comparison. Results show that the proposed approach obtains satisfactory results and outperforms the other comparative approaches.

Broad Attribute Prediction Model With Enhanced Attribute and Feature

For the zero-shot image classification without intersection between training and testing sets, the high-quality representation of image attributes and features plays a key role to improve the classification performance. In order to overcome the limitations related to insufficient attribute and feature expression in zero-shot image classification, we propose a broad attribute prediction model with enhanced attribute and feature (EAF-BAP) based on broad learning and elastic net constraint. Firstly, the EAF-BAP enhances pre-defined attributes by elastic net constraint to obtain hybrid attributes, which effectively improves the finiteness of semantic attributes. Secondly, the enhanced features are constructed by broad learning to increase the discrimination ability of features in different classes. Meanwhile, the broad learning is employed to train multiple attribute classifiers synchronously, which is more efficient compared to traditional support vector machines. Finally, the similarity between predicted attributes and hybrid attributes in testing classes is calculated by Manhattan distance, which is further used to implement image classification. Experiments on both AwA and Shoes datasets show that the proposed EAF-BAP model is capable of improving the accuracy of zero-shot image classification efficiently.

Combustion and Emission Performance of an HCCI Engine Fuelled by n-Heptane/Toluene Blends at a Low-Load Operating Condition

Homogeneous charge compression ignition (HCCI) engine technology offers high fuel efficiency and extra low nitrogen oxide (NOx) and particulate matter (PM) emissions, which makes it a potential alternative combustion mode to conventional diesel engines. A diesel fuel is usually composed of many classes of hydrocarbons among which aromatic compounds have attracted special attention due to their specific combustion characteristics. Understanding the combustion and emission characteristics of different classes of hydrocarbons is crucial for identifying appropriate diesel fuels suitable for HCCI combustion. Toluene is a typical aromatic compound in a diesel fuel. A study of toluene content may provide implications of the effect of aromatic in a diesel fuel on HCCI combustion. In this paper, the combustion and emission performance of an HCCI engine fuelled by n-heptane/toluene blends at a low load operating condition was investigated by experiment and numerical simulation. A modified Cooperative Fuel Research (CFR) engine and a in-house-developed multi-zone model were employed. The engine was operated at the condition of engine speed of 900 rpm, relative air/fuel ratio of 3.5 and without external exhaust gas recirculation. The investigated fuel blends covered a range from pure n-heptane to 70% toluene by volume. Both experimental and numerical results showed that an increase in toluene fraction in the fuel blend retarded combustion phasing. As a result, the optimal compression ratio, at which thermal efficiency reached its maximum for a fuel blend, increased with increasing toluene fraction. The maximum thermal efficiency increased as the toluene fraction increased from 0 to 50%, but then decreased with further increasing toluene fraction to higher values. The peak pressure rise rate also increased with increasing toluene fraction at a constant combustion phasing. An increase in toluene fraction resulted in an increase in unburned hydrocarbon emissions but had little effect on NOx emissions.

Functional principles and regulation of molecular chaperones.

To be able to perform their biological function, a protein needs to be correctly folded into its three dimensional structure. The protein folding process is spontaneous and does not require the input of energy. However, in the crowded cellular environment where there is high risk of inter-molecular interactions that may lead to protein molecules sticking to each other, hence forming aggregates, protein folding is assisted. Cells have evolved robust machinery called molecular chaperones to deal with the protein folding problem and to maintain proteins in their functional state. Molecular chaperones promote efficient folding of newly synthesized proteins, prevent their aggregation and ensure protein homeostasis in cells. There are different classes of molecular chaperones functioning in a complex interplay. In this review, we discuss the principal characteristics of different classes of molecular chaperones, their structure-function relationships, their mode of regulation and their involvement in human disorders.

Data-Adaptive Spatiotemporal ERP Cleaning for Single-Trial BCI Implementation.

This paper presents a data-adaptive approach to enhance the discriminative information of event-related potential (ERP) for the implementation of a brain-computer interface (BCI). The use of single-trial ERP in a real-time BCI application is challenging, due to its inherent noise contamination. Usually, multiple-trial ERPs are averaged to derive discriminative features of different classes by reducing their noise effects. Time-domain filtering is implemented here using an array wavelet transform. Sometimes, several channels can carry the signals, which are irrelevant to actual EPR information against the respective stimuli. A spatial filtering method based on clustering is introduced, to suppress such channels if any. Hence, the single-trial ERP is filtered in both the spatial and temporal domains to improve its discriminative features. The spatial-temporal discriminate analysis is employed to derive the features leading to the performance of target and non-target classification by using linear discriminant analysis. The proposed method is validated using a data set recorded from our experiments. The experimental results show that the performance of the proposed method is superior to that of the recently developed algorithms for single-trial ERP classification.

Prediction of DNase I hypersensitive sites in plant genome using multiple modes of pseudo components

DNase I hypersensitive sites (DHSs) are accessible chromatin zones hypersensitive to DNase I endonucleases in plant genome. DHSs have been used as markers for the presence of transcriptional regulatory elements. It is an important complement to develop computational methods to identify DHSs for discovering potential regulatory elements. To the best of our knowledge, several machine learning approaches have been proposed for the DHSs prediction, but there is still room for improvements. In this work, a new predictor called pDHS-WE was proposed for prediction of DHSs in plant genome by using weighted ensemble learning framework. Here, five classes of heterogeneous features were used to represent the sequences. Five random forest (RF) operators were constructed based on these five classes of features. The proposed pDHS-WE was formed by fusing the five individual RF classifiers into an ensemble predictor. Genetic algorithm was employed to obtain the weights of different classes of features. In the experiments, pDHS-WE obtained accuracy of 88.5%, sensitivity of 89.1%, specificity of 88.0%, and AUC of 0.958, which was more than 2.7%, 2%, 3.5% and 2.6% higher than state-of-the-art methods, respectively. The results suggested that pDHS-WE may become a useful tool for transcriptional regulatory elements analysis in plant genome.

Moving Target Indication Using Deep Convolutional Neural Network

Existing moving target indication (MTI) methods are limited either by the sufficient and homogeneous secondary data (SD) condition or by the high signal-to-clutter-ratio (SCR) requirement. In practice, such constraints are hard to be satisfied. In this paper, we propose a novel deep convolutional neural network (CNN)-based method for the MTI (CNN-MTI) to overcome the limitations of these methods. In the proposed method, the CNN architecture is deliberately designed, and the training data set is greatly augmented to ensure the CNN be deep and well-trained; in addition, the SD only act as the background interference during the augmentation of the training data set. These procedures ensure that the effective features of different classes of moving targets can be extracted, so that good MTI performance can be achieved without the requirement of the sufficient and homogeneous SD or the high SCR condition. Simulation results from the synthetic data and the experimental data demonstrate the validity and the robustness of the CNN-MTI with limited SD support in non-homogeneous and low SCR environment.

Comparative Study on BBA Determination Using Different Distances of Interval Numbers

Dempster-Shafer theory (DST) is an important theory for information fusion. However, in DST how to determinate the basic belief assignment (BBA) is still an open issue. The interval number based BBA determination method is simple and effective, where the features of different classes’ samples are modeled using the interval numbers, i.e., an interval number model is constructed for each focal element.

Phylogenetic analysis of the diacylglycerol kinase family of proteins and identification of multiple highly-specific conserved inserts and deletions within the catalytic domain that are distinctive characteristics of different classes of DGK homologs.

Diacylglycerol kinase (DGK) family of proteins, which phosphorylates diacylglycerol into phosphatidic acid, play important role in controlling diverse cellular processes in eukaryotic organisms. Most vertebrate species contain 10 different DGK isozymes, which are grouped into 5 different classes based on the presence or absence of specific functional domains. However, the relationships among different DGK isozymes or how they have evolved from a common ancestor is unclear. The catalytic domain constitutes the single largest sequence element within the DGK proteins that is commonly and uniquely shared by all family members, but there is limited understanding of the overall function of this domain. In this work, we have used the catalytic domain sequences to construct a phylogenetic tree for the DGK family members from representatives of the main vertebrate classes and have also examined the distributions of various DGK isozymes in eukaryotic phyla. In a tree based on catalytic domain sequences, the DGK homologs belonging to different classes formed strongly supported clusters which were separated by long branches, and the different isozymes within each class also generally formed monophyletic groupings. Further, our analysis of the sequence alignments of catalytic domains has identified >10 novel sequence signatures consisting of conserved signature indels (inserts or deletions, CSIs) that are distinctive characteristics of either particular classes of DGK isozymes, or are commonly shared by members of two or more classes of DGK isozymes. The conserved indels in protein sequences are known to play important functional roles in the proteins/organisms where they are found. Thus, our identification of multiple highly specific CSIs that are distinguishing characteristics of different classes of DGK homologs points to the existence of important differences in the catalytic domain function among the DGK isozymes. The identified CSIs in conjunction with the results of blast searches on species distribution of DGK isozymes also provide useful insights into the evolutionary relationships among the DGK family of proteins.

Steroid bioaccumulation profiles in typical freshwater aquaculture environments of South China and their human health risks via fish consumption

More attention was previously paid to adverse effects of steroids on aquatic organisms and their ecological risks to the aquatic environment. So far, little information has been reported on the bioaccumulative characteristics of different classes of steroids in cultured fish tissues. The present study for the first time provided a comprehensive analysis of the occurrence, bioaccumulation, and global consumers’ health risks via fish consumption of androgens, glucocorticoids and progestanges in typical freshwater cultured farms in South China. The numbers and total concentrations of steroids detected in the tissues of five common species of the cultured fish were in the order of plasma > bile > liver > muscle and plasma > bile, muscle > liver, respectively. The field bioaccumulation factors for the detected synthetic steroids ranged from 450 to 97,000 in bile, 450 to 65,000 in plasma, 2900 to 16,000 in liver, and 42 to 2600 in muscle of fish, respectively. This data suggests that steroids are bioaccumulative in fish tissues. Mostly important, 4-androstene-3,17-dione (AED) and cortisone (CRN) were found to be reliable chemical indicators to predict the levels of steroids in plasma and muscle of the inter-species cultured fish, respectively. Furthermore, the maximum hazard quotients (HQs) of testosterone and progesterone were 5.8 × 10−4 and 9.9 × 10−5, suggesting that human health risks were negligible via ingestion of the steroids-contaminated fish.

A new regularized restricted Boltzmann machine based on class preserving

It is known that an Restricted Boltzmann machine (RBM) can be used as a feature extractor to automatically extract data features in a completely unsupervised learning manner. In this paper, we develop a new regularized RBM by adding the class information, referred to as class preserving RBM (CPr-RBM). Specifically, we impose two constraints on RBM to make the class information clearly reflected in extracted features. One constraint can decrease the distance between the features of the same class and the other one can increase the distance between the features of different classes. The two constraints introduce class information to RBM and make the extracted features contain more category information which contributes to a better classification result. Experiments are conducted on MNIST dataset and 20-newgroup dataset, which show that CPr-RBM learns more discriminate representations and outperforms other related state-of-the-art models in dealing with classification problems.

Intra-night optical variability characteristics of different classes of narrow-line Seyfert 1 galaxies

In a first systematic effort to characterize the intra-night optical variability (INOV) of different classes of narrow line Seyfert 1 (NLSy1) galaxies, we have carried out observations on a sample of radio-loud (RL) and radio-quiet (RQ) NLSy1 galaxies. The RL-NLSy1 galaxies are further divided into {\gamma}-ray loud (GL) and {\gamma}-ray quiet (GQ) NLSy1 galaxies. Our sample consists of four sets, each set consisting of a RQ-NLSy1, a GQ-NLSy1 and a GL-NLSy1 galaxy, closely matched in redshift and optical luminosity. Our observations on both RQ and GQ-NLSy1 galaxies consist of a total of 19 nights, whereas the data for GL-NLSy1 galaxies (18 nights) were taken from literature published earlier by us. This enabled us to do a comparison of the duty cycle (DC) of different classes of NLSy1 galaxies. Using power-enhanced F-test, with a variability threshold of 1%, we find DCs of about 55%, 39% and 0% for GL, GQ and RQ-NLSy1 galaxies respectively. The high DC and large amplitude of INOV (24.0 +/- 13.7%) shown by GL-NLSy1 galaxies relative to the other two classes might be due to their inner aligned relativistic jets having large bulk Lorentz factors. The null DC of RQ-NLSy1 galaxies could mean the presence of low power and/or largely misaligned jets in them. However, dividing RL-NLSy1 galaxies into low and high optical polarization sources, we find that sources with large polarization show somewhat higher DCs (69%) and amplitudes (29%) compared to those with low polarization. This points to a possible link between INOV and optical polarization.

Boosted Jet Tagging with Jet-Images and Deep Neural Networks

Building on the jet-image based representation of high energy jets, we develop computer vision based techniques for jet tagging through the use of deep neural networks. Jet-images enabled the connection between jet substructure and tagging with the fields of computer vision and image processing. We show how applying such techniques using deep neural networks can improve the performance to identify highly boosted W bosons with respect to state-of-the-art substructure methods. In addition, we explore new ways to extract and visualize the discriminating features of different classes of jets, adding a new capability to understand the physics within jets and to design more powerful jet tagging methods.

CloudForest: A Scalable and Efficient Random Forest Implementation for Biological Data.

Random Forest has become a standard data analysis tool in computational biology. However, extensions to existing implementations are often necessary to handle the complexity of biological datasets and their associated research questions. The growing size of these datasets requires high performance implementations. We describe CloudForest, a Random Forest package written in Go, which is particularly well suited for large, heterogeneous, genetic and biomedical datasets. CloudForest includes several extensions, such as dealing with unbalanced classes and missing values. Its flexible design enables users to easily implement additional extensions. CloudForest achieves fast running times by effective use of the CPU cache, optimizing for different classes of features and efficiently multi-threading. https://github.com/ilyalab/CloudForest.