基于最近邻方法的类星体与正常星系光谱分类

发布时间：2019-05-23 00:38

【摘要】：随着高质量CCD传感器技术的日渐成熟与广泛应用,以及许多大型巡天计划的相继实施,天体数据量极大,因此天体观测数据的自动识别、分析问题首当其冲。文章在原始测量空间使用最近邻方法(NN)研究了正常星系与类星体光谱的识别问题。正常星系和类星体属于河外天体,一般距离地球较远,其观测光谱会受到许多干扰,所以这两类天体光谱的分类在光谱自动识别研究中具有一定的代表性。同时,采用的NN是模式识别和数据挖掘方面的基准性方法,在许多新方法的评估中,往往以NN方法的性能作为比较对象。从实用价值来说,研究表明,NN方法的类星体和正常星系光谱识别率与文献中复杂方法的最好结果相当,但该文方法不需要进行分类器的训练,利于实时进行增量式学习和并行实现,这对海量光谱数据的快速处理有重要意义。因此,该研究具有重要的理论参考意义和一定的实用价值。
[Abstract]:With the maturity and wide application of high quality CCD sensor technology, as well as the implementation of many large-scale sky patrol plans, the amount of celestial data is very large, so the automatic identification and analysis of celestial observation data bear the brunt of the problem. In this paper, the nearest neighbor method (NN) is used to study the spectral recognition of normal galaxy and quasar in the original measurement space. Normal galaxy and quasar belong to extragalactic objects, which are generally far from the earth, and their observed spectra will be interfered with by a lot of interference, so the classification of the spectra of these two kinds of celestial bodies is representative in the study of automatic spectral recognition. At the same time, the NN is the benchmark method in pattern recognition and data mining. In the evaluation of many new methods, the performance of NN method is often taken as the comparison object. In terms of practical value, the results show that the spectral recognition rate of quasar and normal galaxy in NN method is similar to that of complex methods in the literature, but this method does not need the training of classifiers. It is beneficial to incremental learning and parallel implementation in real time, which is of great significance to the rapid processing of massive spectral data. Therefore, the study has important theoretical reference significance and certain practical value.
【作者单位】：华南师范大学数学科学学院;潍坊教育学院会计与统计学院;
【基金】：国家自然科学基金项目(61075033,60805028)资助
【分类号】：P152

【参考文献】