傅立叶级数+机器学习法双联预测小行星分类  

Asteroid Classification Based on Joint Fourier Series and Machine Learning

作　　者：徐苑景 XU Yuan-jing(State Key Laboratory of Lunar and Planetary Science,Macao University of Science and Technology,Macao 999078,China)

机构地区：[1]澳门科技大学月球与行星科学国家重点实验室,澳门999078

出　　处：《软件导刊》2020年第11期42-45,共4页Software Guide

摘　　要：利用傅立叶级数对散点小行星光变数据进行曲线拟合,以获得小行星表示参数,选取傅立叶级数为6的拟合方式对单体或双体小行星光变曲线进行分类,并用机器学习算法中的SVM和决策树建立预测模型。检验结果表明,SVM模型对单体和双体小行星的预测正确率达到95%,相较于决策树正确率提高了10%,为从小行星实际观测数据直接推测双体小行星潜在相关应用提供了参考。In this article,the Fourier series is used to fit the light change data of scattered asteroids,so as to obtain the representation parameters of the asteroids.A fitting method is choser with a Fourier series of 6 to classify the light curve of single or binary asteroids,and SVM and decision tree in machine learning algorithm are used to build prediction model.The result shows that the accuracy rate of SVM model predicting single and binary asteroids is 95%,which is 10%higher than that of decision tree.This method can be applied to provide reference for directly estimating the potential twin asteroids from the actual observation data of the asteroids.

关 键 词：傅立叶级数 机器学习 预测 小行星 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]