Discovering gravitationally lensed quasars using supervised machine learning.

使用监督机器学习发现引力透镜类星体。

• 批准号: 1966440
• 金额: --
• 依托单位: University of Cambridge
• 依托单位国家: 英国
• 项目类别: Studentship
• 财政年份: 2017
• 资助国家: 英国
• 起止时间: 2017 至 无数据
• 项目状态: 已结题
• 来源: https://gtr.ukri.org/projects?ref=studentship-1966440
• 关键词: Discovering; gravitationally; lensed; quasars; using

We will adopt a morphology independent supervised machine learning approach to classify quasars and improve the gravitationally lensed quasar search. A Gaussian Mixture Model (GMM) will first be used to select candidates in DES (Dark Energy Survey) Y3 data using features from DES, VHS and WISE.The Gaia satellite provides high resolution multi-epoch (around 70 epochs) positional measurements in the optical for 1 billion astrophysical sources with a complex data model due to the on-board image analysis. Additional features (10 out of 100) measured by Gaia will be added to the GMM analysis, in April 2018, to improve the lensed quasar selection and classification.

我们将采用形态学无关的监督机器学习方法对类星体进行分类，并改进引力透镜类星体搜索。高斯混合模型（GMM）将首先使用来自DES、VHS和WISE的特征来选择DES（暗能量调查）Y3数据中的候选者。盖亚卫星提供高分辨率多历元（约70历元）的光学位置测量，10亿个天体物理源具有复杂的数据模型，由于机载图像分析。2018年4月，盖亚测量到的其他特征（100个中的10个）将被添加到GMM分析中，以改进透镜类星体的选择和分类。