Date of Award
2023
Document Type
Thesis
Degree Name
Master of Science (MS)
Department
Computer Science
Committee Chair
Timothy Newman
Committee Member
Jacob Hauenstein
Committee Member
Nikolai Pogorelov
Subject(s)
Solar flares--Forecasting, Data mining
Abstract
Data mining techniques that can support the forecasting of solar flares are explored. The main focus is to investigate the effects of flare coincidence on the evaluation of forecasts. Coincident flares are solar flares in a region of the Sun in which another flare has occurred recently, whereas other flares are denoted as non-coincident flares. Dimensionality reduction and classification are two of the types of techniques examined, where binary classification is used to issue yes/no forecasts of major solar flares. Five algorithms are considered as binary classifiers: k-nearest neighbors (k-NN), logistic regression (LR), support vector machines (SVM), random forest classifiers (RFC), and linear discriminant analysis (LDA). The methodologies, experimental results, and analyses of utilizing each algorithm are presented. The LR and RFC offered the highest True Skill Statistic (TSS) results.
Recommended Citation
Hall, Christian W., "Solar flare forecasting : a study of flare coincidence" (2023). Theses. 607.
https://louis.uah.edu/uah-theses/607