共查询到20条相似文献，搜索用时 531 毫秒
1.
机器学习在当今诸多领域已经取得了巨大的成功,但是机器学习的预测效果往往依赖于具体问题.集成学习通过综合多个基分类器来预测结果,因此,其适应各种场景的能力较强,分类准确率较高.基于斯隆数字巡天(Sloan Digital Sky Survey,SDSS)计划恒星/星系中最暗源星等集分类正确率低的问题,提出一种基于Stacking集成学习的恒星/星系分类算法.从SDSS-DR7(SDSS Data Release 7)中获取完整的测光数据集,并根据星等值划分为亮源星等集、暗源星等集和最暗源星等集.仅针对分类较为复杂且困难的最暗源星等集展开分类研究.首先,对最暗源星等集使用10折嵌套交叉验证,然后使用支持向量机(Support Vector Machine,SVM)、随机森林(Random Forest,RF)、XGBoost(eXtreme Gradient Boosting)等算法建立基分类器模型;使用梯度提升树(Gradient Boosting Decision Tree,GBDT)作为元分类器模型.最后,使用基于星系的分类正确率等指标,与功能树(Function Tree,FT)、SVM、RF、GBDT、XGBoost、堆叠降噪自编码(Stacked Denoising AutoEncoders,SDAE)、深度置信网络(Deep Belief Network,DBN)、深度感知决策树(Deep Perception Decision Tree,DPDT)等模型进行分类结果对比分析.实验结果表明,Stacking集成学习模型在最暗源星等集分类中要比FT算法的星系分类正确率提高了将近10%.同其他传统的机器学习算法、较强的提升算法、深度学习算法相比,Stacking集成学习模型也有较大的提升. 相似文献
2.
Observations of present and future X‐ray telescopes include a large number of ipitous sources of unknown types. They are a rich source of knowledge about X‐ray dominated astronomical objects, their distribution, and their evolution. The large number of these sources does not permit their individual spectroscopical follow‐up and classification. Here we use Chandra Multi‐Wavelength public data to investigate a number of statistical algorithms for classification of X‐ray sources with optical imaging follow‐up. We show that up to statistical uncertainties, each class of X‐ray sources has specific photometric characteristics that can be used for its classification. We assess the relative and absolute performance of classification methods and measured features by comparing the behaviour of physical quantities for statistically classified objects with what is obtained from spectroscopy. We find that among methods we have studied, multi‐dimensional probability distribution is the best for both classifying source type and redshift, but it needs a sufficiently large input (learning) data set. In absence of such data, a mixture of various methods can give a better final result.We discuss some of potential applications of the statistical classification and the enhancement of information obtained in this way. We also assess the effect of classification methods and input data set on the astronomical conclusions such as distribution and properties of X‐ray selected sources. (© 2008 WILEY‐VCH Verlag GmbH & Co. KGaA, Weinheim) 相似文献
3.
《New Astronomy》2017
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). 相似文献
4.
We study the machine learning method for classifying the basic shape of space debris in both simulated and observed data experiments, where light curves are used as the input features. In the dataset for training and testing, simulated light curves are derived from four types of debris within different shapes and materials. Observed light curves are extracted from Mini-Mega TORTORA (MMT) database which is a publicly accessible source of space object photometric records. The experiments employ the deep convolutional neural network, make comparisons with other machine learning algorithms, and the results show CNN (Convolutional Neural Network) is better. In simulational experiments, both types of cylinder can be distinguished perfectly, and two other types of satellite have around 90% probability to be classified. Rockets and defunct satellites can achieve 99% success rate in binary classification, but in further sub-classes classifications, the rate becomes relatively lower. 相似文献
5.
将未编目的空间碎片正确分类是空间态势感知的重要组成部分. 基于光变曲线, 通过仿真和实测实验, 探讨了空间碎片基本类型的机器学习分类方法. 在数据集中的仿真光变来自形状或材料不同的4类碎片, 实测光变从Mini-Mega TORTORA (MMT)数据库中提取, 实验以深度神经网络作为分类模型, 并和其他机器学习方法进行了比较. 结果显示深度卷积网络优于其他算法, 在仿真实验中对不同材料的圆柱体都能准确识别, 对其余两类卫星的识别率在90%左右; 实测实验中对火箭体和失效卫星的2分类准确率超过99%, 然而在进一步的型号/平台分类中, 准确率有所降低. 相似文献
6.
We introduced a decision tree method called Random Forests for multiwavelength data classification. The data were adopted from different databases, including the Sloan Digital Sky Survey (SDSS) Data Release five, USNO, FIRST and ROSAT.We then studied the discrimination of quasars from stars and the classification of quasars,stars and galaxies with the sample from optical and radio bands and with that from optical and X-ray bands. Moreover, feature selection and feature weighting based on Random Forests were investigated. The performances based on different input patterns were compared. The experimental results show that the random forest method is an effective method for astronomical object classification and can be applied to other classification problems faced in astronomy. In addition, Random Forests will show its superiorities due to its own merits, e.g. classification, feature selection, feature weighting as well as outlier detection. 相似文献
7.
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． 相似文献
8.
Liu Zhong-bao 《Journal of Astrophysics and Astronomy》2016,37(2):12
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. 相似文献
9.
10.
11.
Ayesha Anjum Mousumi Das Jayant Murthy S. B. Gudennavar Rajesh Gopal S. G. Bubbly 《Journal of Astrophysics and Astronomy》2018,39(5):61
We have classified a sample of 37,492 objects from SDSS into QSOs, galaxies and stars using photometric data over five wave bands (u, g, r, i and z) and UV GALEX data over two wave bands (near-UV and far-UV) based on a template fitting method. The advantage of this method of classification is that it does not require any spectroscopic data and hence the objects for which spectroscopic data is not available can also be studied using this technique. In this study, we have found that our method is consistent by spectroscopic methods given that their UV information is available. Our study shows that the UV colours are especially important for separating quasars and stars, as well as spiral and starburst galaxies. Thus it is evident that the UV bands play a crucial role in the classification and characterization of astronomical objects that emit over a wide range of wavelengths, but especially for those that are bright at UV. We have achieved the efficiency of 89% for the QSOs, 63% for the galaxies and 84% for the stars. This classification is also found to be in agreement with the emission line diagnostic diagrams. 相似文献
12.
G. A. Tsirogiannis E. A. Perdios V. V. Markellos 《Celestial Mechanics and Dynamical Astronomy》2009,103(1):49-78
We present an improved grid search method for the global computation of periodic orbits in model problems of Dynamics, and
the classification of these orbits into families. The method concerns symmetric periodic orbits in problems of two degrees
of freedom with a conserved quantity, and is applied here to problems of Celestial Mechanics. It consists of two main phases;
a global sampling technique in a two-dimensional space of initial conditions and a data processing procedure for the classification
(clustering) of the periodic orbits into families characterized by continuous evolution of the orbital parameters of member
orbits. The method is tested by using it to recompute known results. It is then applied with advantage to the determination
of the branch families of the family f of retrograde satellites in Hill’s Lunar problem, and to the determination of irregular families of periodic orbits in a
perturbed Hill problem, a species of families which are difficult to find by continuation methods.
相似文献
13.
In this paper we propose a particle classification system for the imaging calorimeter of the PAMELA satellite-borne experiment. The system consist of three main processing phases. First, a segmentation of the whole signal detected by the calorimeter is performed to select a Region of Interest (RoI); this step allows to retain bounded and space invariant portions of data for the following analysis. In the next step, the RoIs are characterized by means of nine discriminating variables, which measure event properties useful for the classification. The third phase (the classification step) relies on two different supervised algorithms, Artificial Neural Networks and Support Vector Machines. The system was tested with a large simulated data set, composed by 40 GeV/c momentum electrons and protons. Moreover, in order to study the classification power of the calorimeter for experimental data, we have also used biased simulated data. A proton contamination in the range 10−4–10−5 at an electron efficiency greater than 95% was obtained. The results are adequate for the PAMELA imaging calorimeter and show that the approach to the classification based on soft computing techniques is complementary to the traditional analysis performed using optimized cascade cuts on different variables. 相似文献
14.
《New Astronomy》2017
Stellar spectral classification is one of the most fundamental tasks in survey astronomy. Many automated classification methods have been applied to spectral data. However, their main limitation is that the model parameters must be tuned repeatedly to deal with different data sets. In this paper, we utilize the Bayesian support vector machines (BSVM) to classify the spectral subclass data. Based on Gibbs sampling, BSVM can infer all model parameters adaptively according to different data sets, which allows us to circumvent the time-consuming cross validation for penalty parameter. We explored different normalization methods for stellar spectral data, and the best one has been suggested in this study. Finally, experimental results on several stellar spectral subclass classification problems show that the BSVM model not only possesses good adaptability but also provides better prediction performance than traditional methods. 相似文献
15.
《天文和天体物理学研究(英文版)》2020,(9)
We have determined new statistical relations to estimate the fundamental atmospheric parameters of effective temperature and surface gravity, using MK spectral classification, and vice versa. The relations were constructed based on the published calibration tables(for main sequence stars) and observational data from stellar spectral atlases(for giants and supergiants). These new relations were applied to field giants with known atmospheric parameters, and the results of the comparison of our estimations with available spectral classification have been quite satisfactory. 相似文献
16.
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. 相似文献
17.
将基于多重信号分类的MUSIC谱估计算法引入BL Lac天体光变周期分析中.给出了MUSIC算法的基本原理,利用模拟信号检测了算法的频谱分辨率.从大量文献中收集了BL Lac天体S5 0716+714光学V、R、I 3个波段从1994年到2008年的有效观测数据,用MUSIC算法和平均周期图算法分别计算了它们的光变周期,发现存在两个主要光变周期:一个是(3.33±0.08)yr的周期,另一个是(1.24±0.01)yr的周期.对这两种算法的周期估计性能进行了比较,结果表明,MUSIC谱估计算法对样本长度要求较低,具有良好的分辨特性和抗噪声能力,能提高在样本长度较短情况下光变周期分析的准确性. 相似文献
18.
《Chinese Astronomy and Astrophysics》2019,43(4):539-548
Machine learning has achieved great success in many areas today. The lifting algorithm has a strong ability to adapt to various scenarios with a high accuracy, and has played a great role in many fields. But in astronomy, the application of lifting algorithms is still rare. In response to the low classification accuracy of the dark star/galaxy source set in the Sloan Digital Sky Survey (SDSS), a new research result of machine learning, eXtreme Gradient Boosting (XGBoost), has been introduced. The complete photometric data set is obtained from the SDSS-DR7, and divided into a bright source set and a dark source set according to the star magnitude. Firstly, the ten-fold cross-validation method is used for the bright source set and the dark source set respectively, and the XGBoost algorithm is used to establish the star/galaxy classification model. Then, the grid search and other methods are used to adjust the XGBoost parameters. Finally, based on the galaxy classification accuracy and other indicators, the classification results are analyzed, by comparing with the models of function tree (FT), Adaptive boosting (Adaboost), Random Forest (RF), Gradient Boosting Decision Tree (GBDT), Stacked Denoising AutoEncoders (SDAE), and Deep Belief Nets (DBN). The experimental results show that, the XGBoost improves the classification accuracy of galaxies in the dark source classification by nearly 10% as compared to the function tree algorithm, and improves the classification accuracy of sources with the darkest magnitudes in the dark source set by nearly 5% as compared to the function tree algorithm. Compared with other traditional machine learning algorithms and deep neural networks, the XGBoost also has different degrees of improvement. 相似文献
19.
Coryn A. L. Bailer-Jones Mike Irwin & Ted von Hippel 《Monthly notices of the Royal Astronomical Society》1998,298(4):1061-1068
We describe a method for the extraction of spectra from high-dispersion objective prism plates. Our method is a catalogue-driven plate solution approach, making use of the right ascension and declination coordinates for the target objects. In contrast to existing methods of photographic plate reduction, we digitize the entire plate and extract spectra off-line. This approach has the advantages that it can be applied to CCD objective prism images and spectra can be re-extracted (or additional spectra extracted) without having to re-scan the plate. After a brief initial interactive period, the subsequent reduction procedure is completely automatic, resulting in fully reduced, wavelength-justified spectra. We also discuss a method of removing stellar continua using a combination of non-linear filtering algorithms. The method described is used to extract over 12 000 spectra from a set of 92 objective prism plates. These spectra are used in an associated project to develop automated spectral classifiers based on neural networks. 相似文献
20.
One of the most critical points in the detection of cosmic rays by neutron monitors is the correction of the raw data. The data that a detector measures may be distorted by a variety of reasons and the subtraction of these distortions is a prerequisite for processing them further. The final aim of these corrections is to keep only the fluctuations related to the real cosmic-ray intensity. To achieve this, we analyze data from identical neutron monitor detectors which provide a configuration with the ability to exclude the distortions by comparing the counting rate of each detector. Based on this method, a number of effective algorithms have been developed: Median Editor, Median Editor Plus, and Super Editor are some of the algorithms that are being used in the neutron monitor data processing with satisfactory results. In this work, a new approach for the correction of the neutron monitor primary data with a completely different method, based on the use of artificial neural networks, is proposed. A comparison of this method with the algorithms mentioned previously is also presented. 相似文献