發(fā)布時(shí)間：2018-06-19 00:39

  本文選題：恒星 + 基本參數(shù)�。� 參考：《天文學(xué)報(bào)》2016年04期

【摘要】：深度學(xué)習(xí)是當(dāng)前機(jī)器學(xué)習(xí)、模式識(shí)別和人工智能領(lǐng)域中的一項(xiàng)熱點(diǎn)研究技術(shù),非常適用于處理復(fù)雜的大規(guī)模數(shù)據(jù).基于深度學(xué)習(xí)理論構(gòu)建了一個(gè)5層的棧式自編碼深度神經(jīng)網(wǎng)絡(luò),對(duì)恒星大氣物理參數(shù)進(jìn)行自動(dòng)估計(jì),網(wǎng)絡(luò)各層的節(jié)點(diǎn)數(shù)分別為3821-500-100-50-1.使用美國(guó)大型巡天項(xiàng)目Sloan發(fā)布的Sloan Digital Sky Survey(SDSS)實(shí)測(cè)光譜以及由Kurucz的New Opacity Distribution Function(NEWODF)模型得到的理論光譜進(jìn)行了實(shí)驗(yàn)驗(yàn)證,對(duì)有效溫度(Teff)、表面重力加速度(lg g)和金屬豐度([Fe/H])3個(gè)物理參數(shù)進(jìn)行了自動(dòng)估計(jì).結(jié)果表明,棧式自編碼深度神經(jīng)網(wǎng)絡(luò)的估計(jì)精度較好,其中在SDSS數(shù)據(jù)上的平均絕對(duì)誤差分別為:79.95(Teff/K),0.0058(lg(Teff/K)),0.1706(lg(g/(cm·s~(-2)))),0.1294 dex([Fe/H]);在理論數(shù)據(jù)上的平均絕對(duì)誤差分別是:15.34(Teff/K),0.0011(lg(Teff/K)),0.0214(lg(g/(cm·s~(-2)))),0.0121 dex([Fe/H]).
[Abstract]:Deep learning is a hot research technology in the field of machine learning, pattern recognition and artificial intelligence, which is very suitable for dealing with complex large-scale data. Based on the depth learning theory, a five-layer self-coding depth neural network is constructed. The parameters of stellar atmosphere are estimated automatically. The number of nodes in each layer of the network is 3821-500-100-50-1. The measured spectra of Sloan Digital Sky Survey (SDSS) published by Sloan and the theoretical spectra obtained from Kurucz's New Opacity Distribution function ODF model are verified experimentally. Three physical parameters, I. e., effective temperature, surface gravity acceleration (LG) and metal abundance ([Fe / H]), are estimated automatically. 緇撴灉琛ㄦ槑,鏍堝紡鑷紪鐮佹繁搴︾緇忕綉緇滅殑浼拌綺懼害杈冨ソ,鍏朵腑鍦⊿DSS鏁版嵁涓婄殑騫沖潎緇濆璇樊鍒嗗埆涓，

本文編號(hào)：2037532