arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2409.14508astro-ph.HEgr-qc
收藏

评估用于超新星引力波信号分类的机器学习模型

Evaluating Machine Learning Models for Supernova Gravitational Wave Signal Classification

Y. Sultan Abylkairov, Matthew C. Edwards, Daniil Orel, Ayan Mitra, Bekdaulet Shukirgaliyev, Ernazar Abdikamalov

AI总结:

本文通过训练多种机器学习和深度学习模型对旋转核塌缩超新星引力波信号进行分类以探测核物质状态方程,并评估了GREP近似方法对分类精度的影响。

摘要:

我们研究了利用旋转核塌缩超新星产生的引力波(GW)信号来探测核物质状态方程(EOS)的潜力。通过使用不同状态方程的模拟生成引力波信号,我们训练机器学习模型对其进行分类并评估其性能。我们的研究建立在先前工作的基础上,考察了不同机器学习模型、参数和数据预处理技术对分类准确率的影响。我们测试了卷积神经网络和循环神经网络,以及六种经典算法:随机森林、支持向量机、朴素贝叶斯、逻辑回归、$k$近邻和极端梯度提升。除朴素贝叶斯外,所有模型在我们的数据集上均取得了超过90%的准确率。此外,我们评估了使用广义相对论有效势(GREP)近似引力波信号对状态方程分类的影响。我们发现,在GREP数据上训练的模型分类准确率较低。然而,通过峰值信号频率对时间进行归一化,部分补偿了GREP中时间膨胀效应的缺失,使准确率显著提升。尽管如此,准确率仍未超过70%,这表明GREP缺乏状态方程分类所需的精度。最后,我们的研究存在若干局限性,包括未考虑探测器噪声以及仅关注单一前身星质量模型,这些问题将在未来工作中加以解决。

英文摘要:

We investigate the potential of using gravitational wave (GW) signals from rotating core-collapse supernovae to probe the equation of state (EOS) of nuclear matter. By generating GW signals from simulations with various EOSs, we train machine learning models to classify them and evaluate their performance. Our study builds on previous work by examining how different machine learning models, parameters, and data preprocessing techniques impact classification accuracy. We test convolutional and recurrent neural networks, as well as six classical algorithms: random forest, support vector machines, naïve Bayes, logistic regression, $k$-nearest neighbors, and eXtreme gradient boosting. All models, except naïve Bayes, achieve over 90 per cent accuracy on our dataset. Additionally, we assess the impact of approximating the GW signal using the general relativistic effective potential (GREP) on EOS classification. We find that models trained on GREP data exhibit low classification accuracy. However, normalizing time by the peak signal frequency, which partially compensates for the absence of the time dilation effect in GREP, leads to a notable improvement in accuracy. Despite this, the accuracy does not exceed 70 per cent, suggesting that GREP lacks the precision necessary for EOS classification. Finally, our study has several limitations, including the omission of detector noise and the focus on a single progenitor mass model, which will be addressed in future works.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑