arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2402.15874astro-ph.EPastro-ph.IM
收藏

评估分类算法:利用开普勒时间序列数据探测系外行星

Evaluating Classification Algorithms: Exoplanet Detection using Kepler Time Series Data

Fatemeh Fazel Hesar, Bernard Foing

AI总结:

本文系统比较了六种分类算法在开普勒时序数据系外行星探测中的表现,发现随机森林和支持向量机准确率最高且综合性能最优,为后续研究提供了算法选择依据。

摘要:

本研究对使用开普勒任务标记时间序列数据进行系外行星探测的各种分类算法进行了全面评估。研究考察了六种常用算法的性能,即随机森林(Random Forest)、支持向量机(Support Vector Machine)、逻辑回归(Logistic Regression)、K近邻(K-Nearest Neighbors)、朴素贝叶斯(Naive Bayes)和决策树(Decision Tree)。评估过程分析了一个数据集,该数据集包含恒星亮度的时序测量值,并附带指示是否存在系外行星的标签。为评估每种算法在准确识别系外行星方面的有效性,采用了准确率、精确率、召回率和F1分数等性能指标。结果表明,随机森林算法达到了94.2%的最高准确率,紧随其后的是支持向量机,准确率为93.8%。逻辑回归算法的准确率为91.5%,而K近邻、朴素贝叶斯和决策树算法的准确率分别为89.6%、87.3%和85.9%。此外,精确率、召回率和F1分数指标揭示了每个分类器的优势与不足。随机森林算法的精确率为0.92,召回率为0.95,F1分数为0.93,表明其在正确识别正例和负例方面表现均衡。支持向量机也展现出强劲性能,精确率、召回率和F1分数分别为0.91、0.94和0.92。评估表明,随机森林和支持向量机算法非常适合利用开普勒时间序列数据进行系外行星探测。这些发现增进了我们对探测过程的理解,并有助于为未来研究选择合适的算法。

英文摘要:

This study presents a comprehensive evaluation of various classification algorithms used for the detection of exoplanets using labeled time series data from the Kepler mission. The study investigates the performance of six commonly employed algorithms, namely Random Forest, Support Vector Machine, Logistic Regression, K-Nearest Neighbors, Naive Bayes, and Decision Tree. The evaluation process involves analyzing a dataset that consists of time series measurements of star brightness, accompanied by labels indicating the presence or absence of exoplanets. To assess the effectiveness of each algorithm in accurately identifying exoplanets, performance metrics such as accuracy, precision, recall, and F1 score are employed. The results demonstrate that the Random Forest algorithm achieves the highest accuracy of 94.2\%, followed closely by the Support Vector Machine with 93.8 percent accuracy. The Logistic Regression algorithm achieves an accuracy of 91.5 percent, while the K-Nearest Neighbors, Naive Bayes, and Decision Tree algorithms achieve accuracies of 89.6\%, 87.3\%, and 85.9\% respectively. Furthermore, the precision, recall, and F1 score metrics provide insights into the strengths and weaknesses of each classifier. The Random Forest algorithm exhibits a precision of 0.92, recall of 0.95, and F1 score of 0.93, indicating a balanced performance in correctly identifying both positive and negative instances. The Support Vector Machine also demonstrates strong performance with precision, recall, and F1 score values of 0.91, 0.94, and 0.92 respectively. The evaluation demonstrates that Random Forest and Support Vector Machine algorithms are well-suited for exoplanet detection using Kepler time series data. These findings enhance our understanding of the detection process and assist in selecting suitable algorithms for future studies.

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。

↑