Stellar Spectral Classification using Principal Component Analysis and Artificial Neural Networks

A fast and robust method of classifying a library of optical stellar spectra for O to M type stars is presented. The method employs, as tools: (1) principal component analysis (PCA) for reducing the dimensionality of the data and (2) multilayer back propagation network (MBPN) based artificial neural network (ANN) scheme to automate the process of classification. We are able to reduce the dimensionality of the original spectral data to very few components by using PCA and are able to successfully reconstruct the original spectra. A number of NN architectures are used to classify the library of test spectra. Performance of ANN with this reduced dimension shows that the library can be classified to accuracies similar to those achieved by Gulati et al. but with less computational load. Furthermore, the data compression is so efficient that the NN scheme successfully classifies to the desired accuracy for a wide range of architectures. The procedure will greatly improve our capabilities in handling and analysing large spectral data bases of the future.     

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.     

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.     

GEMINI near-infrared spectroscopic observations of young massive stars embedded in molecular clouds

K -band spectra of young stellar candidates in four Southern hemisphere clusters have been obtained with the Gemini Near-Infrared Spectrograph in Gemini South. The clusters are associated with IRAS sources that have colours characteristic of ultracompact H  ii regions. Spectral types were obtained by comparison of the observed spectra with those of a near-infrared (NIR) library; the results include the spectral classification of nine massive stars and seven objects confirmed as background late-type stars. Two of the studied sources have K -band spectra compatible with those characteristic of very hot stars, as inferred from the presence of C  iv , N  iii and N  v emission lines at 2.078, 2.116 and 2.100 μm, respectively. One of them, I16177_IRS1, has a K -band spectrum similar to that of Cyg OB2 7, an O3If* supergiant star. The nebular K -band spectrum of the associated Ultra-Compact (UC) H  ii region shows the s-process [Kr  iii ] and [Se  iv ] high excitation emission lines, previously identified only in planetary nebula. One young stellar object was found in each cluster, associated with either the main IRAS source or a nearby resolved Midecourse Space eXperiment ( MSX ) component, confirming the results obtained from previous NIR photometric surveys. The distances to the stars were derived from their spectral types and previously determined JHK magnitudes; they agree well with the values obtained from the kinematic method, except in the case of IRAS  15408−5356  , for which the spectroscopic distance is about a factor of 2 smaller than the kinematic value.     

Automated classification of sloan digital sky survey (SDSS) stellar spectra using artificial neural networks

Automated techniques have been developed to automate the process of classification of objects or their analysis. The large datasets provided by upcoming spectroscopic surveys with dedicated telescopes urges scientists to use these automated techniques for analysis of such large datasets which are now available to the community. Sloan Digital Sky Survey (SDSS) is one of such surveys releasing massive datasets. We use Probabilistic Neural Network (PNN) for automatic classification of about 5000 SDSS spectra into 158 spectral type of a reference library ranging from O type to M type stars.     

Automated star–galaxy segregation using spectral and integrated band data for TAUVEX/ASTROSAT satellite data pipeline

We employ an Artificial Neural Network (ANN) based technique to develop a pipeline for automated segregation of stars from the galaxies to be observed by Tel-Aviv University Ultra-Violet Experiment (TAUVEX). We use synthetic spectra of stars from UVBLUE library and selected International Ultraviolet Explorer (IUE) low-resolution spectra for galaxies in the ultraviolet (UV) region from 1250 to 3220 Å as the training set and IUE low-resolution spectra for both the stars and the galaxies as the test set. All the data sets have been pre-processed to get band integrated fluxes so as to mimic the observations of the TAUVEX UV imager. We also perform the ANN based segregation scheme using the full length spectral features (which will also be useful for the ASTROSAT mission). Our results suggest that, in the case of the non-availability of full spectral features, the limited band integrated features can be used to segregate the two classes of objects; although the band data classification is less accurate than the full spectral data classification.     

Spectral indices in quantitative spectral classification from stellar libraries

With the use of modern detectors stellar spectral classification libraries have been extended from the photographic regime to the near ℝ at 11000 Å. We have defined new spectral indices within this extended wavelength-range that can be used to determine the luminosity classification for G-K-M stars. An advantage of the new indices, which sample the stellar flux in and out of selected spectral features, is that they are insensitive to catalog differences. This facilitates the use of many catalogs, with varying resolution, different reddening corrections, and calibrations, hence extending the total number of stellar standards available. Furthermore, we demonstrate that the indices can be used to infer absolute magnitudes with good accuracy. The indices should prove useful for analysis of spectra from distant clusters, galaxies, and in particular for problems involving spectral synthesis of stellar populations of galaxies.     

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．     

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.     

Near-IR age-dating of red supergiant-dominated stellar populations

We present new spectral synthesis models for solar metallicity stellar populations, based on a library of stellar spectra that extends across near-IR wavelengths out to 2.4 µm at a resolution approaching 1000. We show that the spectra of massive star clusters in the starburst galaxy M 82 can be reproduced very well with these models. We compare near-IR spectroscopic ages with optical ages, and discuss the main sources of (systematic) errors that still affect those ages.     

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.     

Progress in spectral studies of the FBS blue stellar objects

Spectral observations of 10 FBS blue stellar objects (BSO) with the OHP 1.93-m and BAO 2.6.m telescopes are reported, and overall progress in all the spectral observations and classifications of FBS BSO over 1987–2000 and in the classification of these objects based on all the accessible sources-- in all, 753 out of 1103 objects-- is discussed. Representative slit spectra for the major types of objects are presented and compared with digitized low dispersion spectra from the DFBS. The nature of the FBS objects is examined in terms of advances in the spectral studies. Two-color diagrams are constructed from the SDSS data and are used to find the regions occupied by the various types for further identification of objects of unknown type. Translated from Astrofizika, Vol. 52, No. 1, pp. 85–97 (February 2009).     

A Method of Estimating the [α/Fe] Ratios from Low-resolution Stellar Spectratwo

A method for the determination of [α/Fe] from low-resolution stellar spectra is presented. The proposed scheme includes the following three steps: firstly, the spectrum is decomposed by the multi-scale Haar wavelet, and the high-frequency components are removed to suppress the high-frequency noise; then, based on the correlation of the spectral data component with [α/Fe], the spectral features are selected by the LASSO (Least Absolute Shrinkage and Selection Operator) algorithm; finally, [α/Fe] is measured by the multiple linear regression method based on the MARCS stellar spectrum library. The effectiveness of the method is verified with the low-resolution stellar spectra of ELODIE, SDSS (Sloan Digital Sky Survey), LAMOST (Large Sky Area Multi-Object Fibre Spectroscopic Telescope), and four star clusters. The systematic deviations and accuracies are as follows: (0.04 dex, 0.064 dex) for the 317 ELODIE spectra; (0.16 dex, 0.065 dex) for the 412 SDSS spectra; (0.05 dex, 0.062 dex) for the 1276 LAMOST spectra (with the signal-noise ratio in the g band (SNRG) greater than 20). The averages of [α/Fe] obtained for the likely members of the globular star clusters (M13, M15) and open star clusters (NGC2420, M67) are in agreement with the literature values.     

Rotational broadening and Doppler tomography of the quiescent X-ray nova Centaurus X-4

We present a library of 47 open-cluster integrated spectra, mostly obtained at CASLEO (San Juan, Argentina) in the range  3600 < λ < 7400 Å  , which are made available at CDS. The data are combined with previous spectra to obtain 10 high signal-to-noise ratio basic templates in the young and intermediate-age domains, which are also provided in the library. These Galactic disc templates represent the increased time resolution spectral evolution of a stellar population unit around the Solar metallicity level. The improved signal-to-noise ratio of the present templates with respect to previous template lists, together with their increased time resolution, allowed us to improve the fundamental parameters of some open clusters. The present spectral library will be useful for several astrophysical applications, particularly for population syntheses of star-forming giant galaxies.     

Stellar spectral classification and feature evaluation based on a random forest

With the availability of multi-object spectrometers and the design and operation of some large scale sky surveys, the issue of how to deal with enormous quantities of spectral data efficiently and accurately is becoming more and more important. This work investigates the classification problem of stellar spectra under the assumption that there is no perfect absolute flux calibration, for example, when considering spectra from the Guo Shou Jing Telescope(the Large Sky Area Multi-Object Fiber Spectroscopic Telescope, LAMOST). The proposed scheme consists of the following two procedures: Firstly, a spectrum is normalized based on a 17 th order polynomial fitting; secondly, a random forest(RF) is utilized to classify the stellar spectra. Experiments on four stellar spectral libraries show that the RF has good classification performance. This work also studied the spectral feature evaluation problem based on RF. The evaluation is helpful in understanding the results of the proposed stellar classification scheme and exploring its potential improvements in the future.     

Spectral classification of stars in the second byurakan sky survey. I

Results are given on a spectral classification of 316 stars and objects having a continuous spectrum, selected in the course of the Second Byurakan Spectral Sky Survey. Slit spectra are used, obtained from 1978 to 1994, predominantly on the 6-m telescope of the Special Astrophysical Observatory, Russian Academy of Sciences, with a resolution of 5–15 Å. The sample contains objects in the range of stellar magnitudes 10.12 ≤ m(pg) ≤ 19.5. We found 114 DA, 13 DB, 4 DC, and 1 DQ white dwarfs, 57 sdB and 13 sdO subdwarfs, 12 NHB, 10 cataclysmic variables, 76 stars of late spectral types, 3 spectroscopic binary systems, and 13 objects with a continuous spectrum. Several recordings are given for each of these spectral types.     

Spectral classification of stars based on LAMOST spectra

In this work, we select spectra of stars with high signal-to-noise ratio from LAMOST data and map their MK classes to the spectral features. The equivalent widths of prominent spectral lines, which play a similar role as multi-color photometry, form a clean stellar locus well ordered by MK classes. The advantage of the stellar locus in line indices is that it gives a natural and continuous classification of stars consistent with either broadly used MK classes or stellar astrophysical parameters. We also employ an SVM-based classification algorithm to assign MK classes to LAMOST stellar spectra. We find that the completenesses of the classifications are up to 90% for A and G type stars, but they are down to about 50% for OB and K type stars. About 40% of the OB and K type stars are mis-classified as A and G type stars,respectively. This is likely due to the difference in the spectral features between late B type and early A type stars or between late G and early K type stars being very weak. The relatively poor performance of the automatic MK classification with SVM suggests that the direct use of line indices to classify stars is likely a more preferable choice.     

PopStar I: evolutionary synthesis model description

We present new evolutionary synthesis models for simple stellar populations for a wide range of ages and metallicities. The models are based on the Padova isochrones. The core of the spectral library is provided by the medium resolution Lejeune et al. atmosphere models. These spectra are complemented by Non Local Thermodynamic Equilibrium (NLTE) atmosphere models for hot stars that have an important impact on the stellar cluster's ionizing spectra: O, B and WR stellar spectra at the early ages, and spectra of post asymptotic giant branch stars and planetary nebulae, at intermediate and old ages. At young ages, our models compare well with other existing models, but we find that the inclusion of the nebular continuum, not considered in several other models, significantly reddens the integrated colours of very young stellar populations. This is consistent with the results of spectral synthesis codes particularly devised for the study of starburst galaxies. At intermediate and old ages, the agreement with the literature model is good and, in particular, we reproduce the observed colours of star clusters in Large Magellanic Cloud well. Given the ability to produce good integrated spectra from the far-ultraviolet to the infrared at any age, we consider that our models are particularly suited for the study of high-redshift galaxies. These models are available on the web site http://www.fractal-es.com/SEDmod.htm and also through the Virtual Observatory Tools on the PopStar server.     

使用SOFM方法进行恒星光谱自动分类

ＳＯＦＭ是人工神经网络的非监督学习算法,可以将数据组织到一个特征图上,而保存 大多数原始数据空间的拓扑特征．使用这种方法进行恒星光谱自动分类,分类结果与哈佛 序列十分相似．ＳＯＦＭ方法应该是进行大数量恒星光谱样本在线分类的有用方法,它能 够自动执行,因此可用于处理大数量天体光谱．     

Spectral classification of stars in the second Byurakan Sky Survey. II

The results of the spectra/classification of 310 stars and objects having a continuous spectrum, selected in the course of the Second Byurakan Spectral Sky Survey, are given. Slit spectra were used in the work, which were obtained from 1978 to 1994, mainly on the 6-m telescope of the Special Astrophysical Observatory, Russian Academy of Sciences; some of the spectra were obtained on the 2.6-m telescope of the Byurakan Astrophysical Observatory, National Academy of Sciences of Armenia, and the 4.5-m Multiple-Mirror Telescope (USA), with 5–15 Å resolution. The sample contains objects in the range of stellar magnitudes 11.0 ≤ m(pg) ≤ 19.5. We found 93 DA, 2 DO, 12 DB, and 1 DC white dwarfs, 83 sdB and 20 sdO subdwarfs, 34 NHB variables, 2 cataclysmic variables, 48 stars of late spectral types, 8 spectroscopic binary systems, and 7 objects having a continuous spectrum. Several recordings are given for each of these spectral types.