Machine learning technique for morphological classification of galaxies from the SDSS
Machine learning technique for morphological classification of galaxies from the SDSS
复制标题
DOI：
10.1051/0004-6361/202038981
复制
发表时间：
2017-12
影响因子：
6.5
通讯作者：
I. Vavilova;D. Dobrycheva;M. Vasylenko;A. Elyiv;O. Melnyk;V. Khramtsov
中科院分区：
文献类型：
--
作者：
I. Vavilova;D. Dobrycheva;M. Vasylenko;A. Elyiv;O. Melnyk;V. Khramtsov
上下文机器学习方法是天文任务中根据对象的个体特征对其进行分类的有效工具。其中一个很有前途的实用程序是有关的星系在不同的红移形态分类。目标。我们使用基于光度测量的方法对SDSS数据（1）开发五种监督机器学习技术，并确定其中最有效的用于自动星系形态分类;（2）测试光度测量数据对形态分类的影响;（3）讨论监督机器学习和标记偏差的问题点;（4）应用最佳拟合的机器学习方法从SDSS DR 9中揭示z < 0.1的未知星系形态类型。方法.我们使用了不同的星系分类技术：人类标记，多光度图，朴素贝叶斯，逻辑回归，支持向量机，随机森林，k近邻。结果我们提出了一个二进制的自动形态分类的星系进行人类标记，多测光，和五个监督机器学习方法的结果。我们将其应用于SDSS DR 9的星系样本，其红移为0.02 < z < 0.1，绝对星等为− 24 m < Mr < − 19.4 m。在分析中，我们使用了绝对星等Mu、Mg、Mr、Mi、Mz;颜色指数Mu − Mr、Mg − Mi、Mu − Mg、Mr − Mz;以及中心的反浓度指数R50/R90。我们确定了每种方法预测形态类型的能力，并验证了该方法的准确性对红移，人类标记，形态形状和重叠的不同形态类型的星系具有相同的颜色指数的各种依赖关系。我们发现，基于光度参数训练的监督机器学习方法的形态学比基于公民科学分类器的形态学表现出更少的偏差。结论.支持向量机和随机森林方法与Python中的Scikit-learn软件机器学习库为二元星系形态分类提供了最高的准确性。具体而言，支持向量机的成功率为96.4%（96.1%早期E型和96.9%晚期L型），随机森林的成功率为95.5%（96.7%早期E型和92.8%晚期L型）。对SDSS DR 9中z < 0.1的316 031个未知形态类型的星系样本进行支持向量机分类，得到139 659个E型和176 372个L型。
Context. Machine learning methods are effective tools in astronomical tasks for classifying objects by their individual features. One of the promising utilities is related to the morphological classification of galaxies at different redshifts. Aims. We use the photometry-based approach for the SDSS data (1) to exploit five supervised machine learning techniques and define the most effective among them for the automated galaxy morphological classification; (2) to test the influence of photometry data on morphology classification; (3) to discuss problem points of supervised machine learning and labeling bias; and (4) to apply the best fitting machine learning methods for revealing the unknown morphological types of galaxies from the SDSS DR9 at z < 0.1. Methods. We used different galaxy classification techniques: human labeling, multi-photometry diagrams, naive Bayes, logistic regression, support-vector machine, random forest, k-nearest neighbors. Results. We present the results of a binary automated morphological classification of galaxies conducted by human labeling, multi-photometry, and five supervised machine learning methods. We applied it to the sample of galaxies from the SDSS DR9 with redshifts of 0.02 < z < 0.1 and absolute stellar magnitudes of −24m < Mr < −19.4m. For the analysis we used absolute magnitudes Mu, Mg, Mr, Mi, Mz; color indices Mu − Mr, Mg − Mi, Mu − Mg, Mr − Mz; and the inverse concentration index to the center R50/R90. We determined the ability of each method to predict the morphological type, and verified various dependencies of the method’s accuracy on redshifts, human labeling, morphological shape, and overlap of different morphological types for galaxies with the same color indices. We find that the morphology based on the supervised machine learning methods trained over photometric parameters demonstrates significantly less bias than the morphology based on citizen-science classifiers. Conclusions. The support-vector machine and random forest methods with Scikit-learn software machine learning library in Python provide the highest accuracy for the binary galaxy morphological classification. Specifically, the success rate is 96.4% for support-vector machine (96.1% early E and 96.9% late L types) and 95.5% for random forest (96.7% early E and 92.8% late L types). Applying the support-vector machine for the sample of 316 031 galaxies from the SDSS DR9 at z < 0.1 with unknown morphological types, we found 139 659 E and 176 372 L types among them.