Generation of a supervised classification algorithm for time-series variable stars with an application to the LINEAR dataset

With the advent of digital astronomy, new benefits and new problems have been presented to the modern day astronomer. While data can be captured in a more efficient and accurate manner using digital means, the efficiency of data retrieval has led to an overload of scientific data for processing and storage. This paper will focus on the construction and application of a supervised pattern classification algorithm for the identification of variable stars. Given the reduction of a survey of stars into a standard feature space, the problem of using prior patterns to identify new observed patterns can be reduced to time-tested classification methodologies and algorithms. Such supervised methods, so called because the user trains the algorithms prior to application using patterns with known classes or labels, provide a means to probabilistically determine the estimated class type of new observations. This paper will demonstrate the construction and application of a supervised classification algorithm on variable star data. The classifier is applied to a set of 192,744 LINEAR data points. Of the original samples, 34,451 unique stars were classified with high confidence (high level of probability of being the true class).     

Statistical confirmation of the UV spectral classification system based on IUE Low-Dispersion stellar spectra

In theIUE Low-Dispersion Spectra Reference Atlas (Hecket al., 1984a), a new spectral classification system specific to the UV had to be introduced because of the lack of one-to-one correspondence between the UV and visible ranges. It was elaborated from a classical morphological approach (Jaschek and Jaschek, 1984).This paper presents an independent confirmation of the correctness of this system. A statistical methodology working in a multidimensional parametric space has been applied to variables expressing, as objectively as possible, the information contained in the continuum and the spectral features of a set of stellar low-dispersion IUE spectra. This was done through, on the one hand, an asymmetry coefficient describing the continuum shape and empirically corrected for the interstellar reddening, and, on the other hand, the intensities of sixty objectively selected lines.These line intensities have been weighted in a way we called the variable Procrustean bed method because, contrary to a standard weighting where a variable is weighted in the same way for all the individuals of a sample, the spectral variables were weighted here according to the asymmetry coefficient which varies with the star at hand. The statistical algorithm consisted of a Principal-Component Analysis followed by a Cluster Analysis.The choice of the lines used for the morphological approach in the Atlas is shown to be correct. With respect to the UV classification system introduced in the atlas, the groups constructed by the cluster analysis display good homogeneity and discrimination for spectral types and luminosity classes, especially in the early spectral types which are well represented in the sample used for this study. The UV standard stars can be found in the neighbourhood of the barycenters of the groups. Moreover, the methodology developed here could be used in a later stage to predict UV spectral classifications.Based on observations collected by the International Ultraviolet Explorer (IUE) at the European Space Agency Villafranca Satellite Tracking Station (VILSPA) and on IUE data retrieved from the VILSPA data bank.     

Removing cosmic-ray hits from CCD images in real-time mode by means of an artificial neural network

A feed-forward artificial neural network has been implemented to the problem of removing cosmic-ray hits (CRH) from CCD images. The results of a number of tests demonstrate the effectiveness of this method especially for undersampled stellar profiles. The problem of optimal and low price preparing of training data, which could enable real-time or at least fast post-processing filtering out of CRH is discussed. The training and test ensembles were composed of a number of synthetic stellar profiles involving different S/N ratios and CRH images taken from real data. Certain aspects of the network’s architecture and its training efficiency for different modes of the back-propagation procedure as well as for the pre-process normalization of data have been examined. It is shown that for training set composed of stellar images and CRH at a ratio of 1:2 recognition can reach 99% in the case of stars and 96% for CRH. To determine the extent to which the cognition power of a network trained using an ensemble of circular symmetric stellar profiles of a given radius can be generalised the test data included stellar profiles of different radii, as well as elongated profiles. The goal was to mimic temporal changes in seeing as well as such problems as image defocusing, the lack of isoplanatism and improper sideral tracking of a telescope. The experiments provided us with the conclusion that for S/N > 10 excellent classification property is maintained in cases where the change in the radius of a circular profile is up to 30%, as well as for elongated profiles where the longest dimension is almost double that of the shortest one. Moreover, the generalization capability has been investigated for test images of synthetic pairs of overlapping stars with different distances between components. Almost 99% recognition efficiency was achieved even if the separation was nearly three times the radius of the stellar profile, a case when two stars could be analyzed by appropriate software as separate objects. The example of removal of CRH from real CCD images is presented to give an idea of how an algorithm based on a neural network can work in practice. The result of such an experiment appears fully consistent with the conclusions drawn from the tests made on synthetic data.     

一种新的基于2维傅里叶谱图像的恒星光谱特征提取方法和深度网络分类应用

天体光谱分类是天文学研究的重要内容之一,其关键是从光谱数据中选择和提取对分类识别最有效的特征构建特征空间.提出一种新的基于2维傅里叶谱图像的特征提取方法,并应用于LAMOST (the Large Sky Area Multi-Object Fiber Spectroscopic Telescope)恒星光谱数据的分类研究中.光谱数据来源于LAMOST Data Release 5(DR5),选取30000条F、 G和K型星光谱数据,利用短时傅里叶变换(Short-Time Fourier Transform, STFT)将1维光谱数据变换成2维傅里叶谱图像,对得到的2维傅里叶谱图像采用深度卷积网络模型进行分类,得到的分类准确率是92.90%.实验结果表明通过对LAMOST恒星光谱数据进行STFT可得到光谱的2维傅里叶谱图像,谱图像构成了新的光谱数据特征和特征空间,新的特征对于光谱数据分类是有效的.此方法是对光谱分类的一种全新尝试,对海量天体光谱的分类和挖掘处理有一定的开创意义.     

A multi-layered backpropagation artificial neural network algorithm for UV spectral classification

In this paper we present an application of an artificial neural network model based on a multi-layered backpropagation algorithm for spectral classification of UV data from the International Ultraviolet Explorer (IUE) low dispersion spectra reference atlas. The model used is similar to that of von Hippel et al. (1994), and is found to reduce the classification error as compared to the recently reported results on the same data set (Gulati et al. 1994b). The improved version of the network is much simpler in structure and the training time is reduced by a factor of almost 20. Such networks will prove very useful in efficient classification of large databases Subject headings: neural networks, stellar spectra, classification     

MK标准星的CCD光谱及光谱分类判据向黄红区的扩展

本文提供了１２５颗ＭＫ标准星的ＣＣＤ光谱,光谱型从Ｏ到Ｍ,光度级从Ｖ到Ⅰ,构成较完整的二元分类框架,光谱覆盖范围由传统蓝紫区延伸到黄红区．初步考察和归纳了黄红区适于恒星分类的主要光谱特征和判据．这些结果对于采用相似分辨率的恒星光谱分类工作是非常有用的．     

On the use of logistic regression for stellar classification

We are totally immersed in the Big Data era and reliable algorithms and methods for data classification are instrumental for astronomical research. Random Forest and Support Vector Machines algorithms have become popular over the last few years and they are widely used for different stellar classification problems. In this article, we explore an alternative supervised classification method scarcely exploited in astronomy, Logistic Regression, that has been applied successfully in other scientific areas, particularly biostatistics. We have applied this method in order to derive membership probabilities for potential T Tauri star candidates from ultraviolet-infrared colour-colour diagrams.     

恒星大气物理参量的非参数估计方法

恒星大气物理参量（有效温度、表面重力、化学丰度）是导致恒星光谱差异的主要因素．恒星大气物理参量的自动测量是LAMOST等大规模巡天望远镜所产生的海量天体光谱数据自动处理中一个重要研究内容．针对测量大样本的恒星光谱数据估计每个恒星的大气物理参量，提出了一种基于变窗宽核函数的估计算法：变窗宽算法是对固定窗宽算法的改进，分为3个步骤：（1）将历史恒星光谱数据进行PCA处理，得到光谱的低维特征数据；（2）利用特征数据与其物理参数的对应关系，建立一种变窗宽的非参数估计模型；（3）利用该估计模型，直接计算待测恒星光谱的3个物理参量（有效温度、表面重力、金属丰度）．实验结果表明：该方法与固定窗宽估计模型以及在其他文献中报道的方法相比，具有较高的估计精度和鲁棒性．     

Classification of stellar spectra with SVM based on within-class scatter and between-class scatter

Support Vector Machine (SVM) is a popular data mining technique, and it has been widely applied in astronomical tasks, especially in stellar spectra classification. Since SVM doesn’t take the data distribution into consideration, and therefore, its classification efficiencies can’t be greatly improved. Meanwhile, SVM ignores the internal information of the training dataset, such as the within-class structure and between-class structure. In view of this, we propose a new classification algorithm-SVM based on Within-Class Scatter and Between-Class Scatter (WBS-SVM) in this paper. WBS-SVM tries to find an optimal hyperplane to separate two classes. The difference is that it incorporates minimum within-class scatter and maximum between-class scatter in Linear Discriminant Analysis (LDA) into SVM. These two scatters represent the distributions of the training dataset, and the optimization of WBS-SVM ensures the samples in the same class are as close as possible and the samples in different classes are as far as possible. Experiments on the K-, F-, G-type stellar spectra from Sloan Digital Sky Survey (SDSS), Data Release 8 show that our proposed WBS-SVM can greatly improve the classification accuracies.     

Automated Separation of Stars and Normal Galaxies Based on Statistical Mixture Modeling with RBF Neural Networks

For LAMOST, the largest sky survey program in China, the solution of the problem of automatic discrimination of stars from galaxies by spectra has shown that the results of the PSF test can be significantly refined. However, the problem is made worse when the redshifts of galaxies are not available. We present a new automatic method of star/(normal) galaxy separation, which is based on Statistical Mixture Modeling with Radial Basis Function Neural Networks (SMM-RBFNN). This work is a continuation of our previous one, where active and non-active celestial objects were successfully segregated. By combining the method in this paper and the previous one, stars can now be effectively separated from galaxies and AGNs by their spectra-a major goal of LAMOST, and an indispensable step in any automatic spectrum classification system. In our work, the training set includes standard stellar spectra from Jacoby's spectrum library and simulated galaxy spectra of EO, SO, Sa, Sb types with redshift ranging from 0 to 1     

Elemental abundances in the Milky Way stellar disk(s), bulge,and halo

We present a review of elemental abundances in the Milky Way stellar disk, bulge, and halo with a focus on data derived from high-resolution stellar spectra. These data are fundamental in disentangling the formation history and subsequent evolution of the Milky Way. Information from such data is still limited and confined to narrowly defined stellar samples. The astrometric Gaia satellite will soon be launched by the European Space Agency. Its final data set will revolutionize information on the motions of a billion stars in the Milky Way. This will be complemented by several ground-based observational campaigns, in particular spectroscopic follow-up to study elemental abundances in the stars in detail. Our review shows the very rich and intriguing picture built from rather small and local samples. The Gaia data deserve to be complemented by data of the same high quality that have been collected for the solar neighborhood.     

Mode constraints on nonradial pulsations from polarization data

The distortion of the surface of a hot variable star undergoing nonradial pulsation produces a variable polarization. Resultant changes in the Stokes parameters are here calculated analytically in the linear regime as a function of both the pulsation mode (l, m) and the orientation of the stellar rotation axis. It is shown that all of the necessary stellar atmosphere information may be condensed into one simple integral. The analytic form permits a systematic appraisal of the mode-discriminating ability of polarization observations. A suitable systematic approach is outlined and illustrated with reference to the data for BW Vul.     

Long-duration solar and stellar flares

According to one of the most popular classifications, solar flares may be assigned either to the category of small short-lived events, or to the category of large, long-duration two-ribbon (2-R) flares. Even if such abroad division oversimplifies the flare phenomenon, our knowledge of the characteristics of stellar flares is so poor, that it is worthwhile to investigate the possibility of adopting this classification scheme for stellar flares as well. In particular we will analyze Einstein observations of a long duration flare on EQ Peg to establish whether it might be considered as a stellar analogy of 2-R solar events. To this end we apply to EQ Peg data a reconnection model, developed originally for solar 2-R flares, and conclude that stellar observations are consistent with model predictions, although additional information is required to identify uniquely the physical parameters of the flare region. Application of the model to integrated observations of a 2-R solar flare, for which high spatial resolution data are also available, shows, however, that future integrated observations may allow us to solve the ambiguities of the model and use it as a diagnostic tool for a better understanding of stellar flares.     

Stellar Spectral Classification with Locality Preserving Projections and Support Vector Machine

With the help of computer tools and algorithms, automatic stellar spectral classification has become an area of current interest. The process of stellar spectral classification mainly includes two steps: dimension reduction and classification. As a popular dimensionality reduction technique, Principal Component Analysis (PCA) is widely used in stellar spectra classification. Another dimensionality reduction technique, Locality Preserving Projections (LPP) has not been widely used in astronomy. The advantage of LPP is that it can preserve the local structure of the data after dimensionality reduction. In view of this, we investigate how to apply LPP+SVM in classifying the stellar spectral subclasses. In the comparative experiment, the performance of LPP is compared with PCA. The stellar spectral classification process is composed of the following steps. Firstly, PCA and LPP are respectively applied to reduce the dimension of spectra data. Then, Support Vector Machine (SVM) is used to classify the 4 subclasses of K-type and 3 subclasses of F-type spectra from Sloan Digital Sky Survey (SDSS). Lastly, the performance of LPP+SVM is compared with that of PCA+SVM in stellar spectral classification, and we found that LPP does better than PCA.     

Correcting CCD photometry of stars for seeing effects

Systematic variability in stellar magnitudes, as derived from profile fitting to CCD images, may in some instances be due to variable seeing. It is suggested that this happens in cases where the stars are unresolved pairs, typically with sub-arcsecond separation between the components. It is shown that the fitting of suitable Generalised Additive Models to time series photometry can disentangle intrinsic stellar variability and seeing-induced brightness changes. It is possible that there will be a fixed seeing response associated with a given star which exhibits the effect: estimation of this response from several long photometric runs is demonstrated.     

A Road Map for the Generation of a Near-Infrared Guide Star Catalog for Thirty Meter Telescope Observations

The near-infrared instruments in the upcoming Thirty Meter Telescope (TMT) will be assisted by a multi conjugate Adaptive Optics (AO) system. For the efficient operation of the AO system, during observations, a near-infrared guide star catalog which goes as faint as 22 mag in J_Vega band is essential and such a catalog does not exist. A methodology, based on stellar atmospheric models, to compute the expected near-infrared magnitudes of stellar sources from their optical magnitudes is developed. The method is applied and validated in JHKs bands for a magnitude range of J_Vega 16–22 mag. The methodology is also applied and validated using the reference catalog of PAN STARRS. We verified that the properties of the final PAN STARRS optical catalog will satisfy the requirements of TMT IRGSC and will be one of the potential sources for the generation of the final catalog. In a broader context, this methodology is applicable for the generation of a guide star catalog for any existing/upcoming near-infrared telescopes.     

Stellar Spectral Classification Based on Capsule Network

The rapid development of large-scale sky survey project has produced a large amount of stellar spectral data, which make the automatic classification of stellar spectral data a challenging task. In this paper, we have proposed a stellar spectral classification method based on a capsule network. At first, by using the one-dimensional convolutional network and short-time Fourier transform (STFT), the one-dimensional spectra of the F5, G5, and K5 types selected from the LAMOST Data Release 5 (DR5) are converted into the two-dimensional Fourier spectrum images. Then, the two-dimensional Fourier spectrum images are classified automatically by the capsule network. Because the capsule network can preserve the hierarchical pose relationships among the entities in the image, and it does not need any pooling layers, the experimental results show that the capsule network has a better classification performance, for the classifications of the F5, G5, and K5-type stellar spectra, its classification accuracy is superior to other classification methods.     

NO emissions as observed by SPICAV during stellar occultations

Ultraviolet (UV) nightglow data from the SPICAV instrument (SPectroscopy for the Investigation of the Characteristics of the Atmosphere of Venus) onboard the Venus Express spacecraft, currently in orbit around Venus, are presented. In its extended source mode, SPICAV has shown that the Venus nightglow in the UV contains essentially Lyman-α and Nitric Oxide (NO) emissions. In the stellar mode, when the slit of the spectrometer is removed, an emission is also observed at the limb in addition to the stellar spectrum. A forward model allows us to identify this feature as being an NO emission. Due to radiative recombination of N and O atoms produced on the dayside of Venus, and transported to the nightside, NO nightglow provides important constraints to the Solar-to-Anti Solar thermospheric circulation prevailing above 90 km. The forward model presented here allows us to derive the altitude of the peak of emission of the NO layer, found at 113.5±6 km, as well as its scale height, of 3.4±1 km and its brightness. The latter is found to be very variable with emissions between 19 Kilo-Rayleigh (kR) and 540 kR. In addition, the NO nightglow is sometimes very patchy, as we are able to observe two distinct emission zones in the field of view. Finally, systematic extraction of this emission from stellar occultations extends the database of the NO emission already reported elsewhere using limb observations.     

Multivariate study of dynamically hot stellar systems: Clues to the origin of ultra compact and ultra faint dwarfs

A multivariate classification has been performed for a large sample of dynamically hot stellar systems comprising globular clusters to giant ellipticals, in quest of the formation theory of ultra compact dwarf galaxies (UCDs). For this K means cluster analysis is carried out together with the optimum criterion (Sugar et al., 2003) with respect to three parameters, logarithm of stellar mass, logarithm of effective radius and stellar mass to light ratio. The present data set has been taken from Misgeld and Hilker (2011). We found five groups MK1–MK5. These are predominated by giant ellipticals (gEs), faint dwarf ellipticals (dEs), globular clusters (GCs), massive compact objects (UCDs and nuclei of dE,Ns) and bright dwarf ellipticals respectively. Almost all UCDs are found either in MK3 or MK4. The fraction is roughly 50%–50% between MK3 and MK4. Comparable fraction of UCDs share properties either with normal GCs or with nuclei of dE,N. This adds a quantitative constraint to the long discussed hypothesis that UCDs may be formed either as massive globular clusters or have an origin similar to nuclei of dwarf galaxies. We finally find that for our clustering test in mass-size-stellar M/L ratios, ultra faint dwarf galaxies are attributed to globular cluster group (MK3) and not to the dwarf galaxy group (MK2). This highlights that there is no clear cut morphological distinction between extended star clusters and ultra faint dwarfs. These groups are highly consistent with the groups found in a previous classification for a smaller sample and completely different set of parameters.     

基于胶囊网络的恒星光谱分类研究

大型巡天项目的快速发展,产生大量的恒星光谱数据,也使得实现恒星光谱数据的自动分类成为一项具有挑战性的工作.提出一种新的基于胶囊网络的恒星光谱分类方法,首先利用1维卷积网络和短时傅里叶变换将来源于LAMOST(Large Sky Area Multi-Object Fiber Spectroscopy Telescope)Data Release 5(DR5)的F5、G5、K5型1维恒星光谱转化成2维傅里叶谱图像,再通过胶囊网络对2维谱图像进行自动分类.由于胶囊网络具有保留图像中实体之间的分层位姿关系和无需池化层的优点,实验结果表明:胶囊网络具有较好的分类性能,对于F5、G5、K5型恒星光谱的分类,准确率优于其他分类方法.