应用数据挖掘技术的短期太阳耀斑预报模型

发布时间：2018-10-09 16:03

【摘要】：为了进一步探讨太阳耀斑与太阳黑子参量的关系,本文采集了大规模的活动区黑子数据,统计其与耀斑发生的产率关系,应用得到的拟和公式对原始数据计算得到规范化后的数据集.在此基础上使用数据挖掘技术对黑子耀斑数据建立决策树模型和建立分类规则,具体描述了黑子数据和太阳耀斑之间的相关性.最后应用这两种技术对活动区未来48h是否爆发耀斑给出了预报,预报结果具有较高的准确率和较低的虚报率.
[Abstract]:In order to further study the relationship between solar flares and sunspot parameters, the large scale active region sunspot data are collected, and the relation between solar flares and the yield of solar flares is calculated. The normalized data set is obtained by applying the quasi-sum formula to the calculation of the original data. On this basis, the decision tree model and classification rules of sunspot flare data are established by using data mining technology, and the correlation between sunspot data and solar flare is described in detail. Finally, the two techniques are used to predict whether flares will erupt in the active region in the next 48 hours. The prediction results have higher accuracy and lower false report rate.
【作者单位】：北京物资学院;中国科学院空间科学与应用研究中心;
【基金】：国家自然科学基金(编号:10973020) 北京市属高等学校人才强教计划资助项目(编号:PHR200906210) 北京市教育委员会科研基地建设项目(编号:WYJD200902);北京市教育委员会科技计划项目(编号:KM200810037001)资助
【分类号】：TP311.13;P182.52

【共引文献】