| 摘要: | 我開發了一個多算法機器學習模型，用以分析光變曲線來探測恆星閃焰。我定義出的四種閃焰光變曲線特徵作為訓練指標，來訓練深度神經網絡 (Deep Neural Network, DNN)、隨機森林 (Random Forest, RF)和XGBoost演算法的模型。這些模型性能通過準確率、精確率、召回率和F1分數進行評估，大多超過94%。我利用此模型分析以前已發表的閃焰文獻的資料，在與文獻中的TESS M型矮星閃焰識別結果進行比對驗證後，我的模型成功重新探測到了超過92%的已知閃焰，同時還檢測到約2000個文獻中未發現的小型事件。這測試證明了我的模型有比以往文獻中的閃焰探測法有更高的靈敏度。經過處理130萬個光變曲線後，我的模型識別出近18,000顆閃焰恆星和250,000個閃焰。我將這些發現統整成一個大型目錄並發表公開，該目錄記錄了閃焰和與閃焰星的特性。我的結果表明總閃焰能量和閃焰振幅與顏色之間存在強相關性，與先前的研究一致。我也對閃焰頻率分佈進行分析，與先前研究不同的是，這一次我評估了由於低振幅事件檢測不完全性而引起的頻率誤差，進而改進了閃焰頻率分布的幂律斜率。我測定了約120,000顆恆星的自轉週期，從而揭示了自轉週期與閃焰活動之間的關係。我發現，恆星的自轉週期可明確分別出閃焰能量飽和區以及未飽和區，也能分別出coronal X射線發射飽和以及未飽和區，這表明了恆星的Coronal heating機制與閃焰能量高度相關。我還發現，在早期型和未飽和恆星中，X射線發射隨著閃焰光度增加得更快，表明這些天體中有更高效的coronal heating。另外，我在一些白矮星 (white dwarfs)和熱亞矮星(hot subdwarfs)中野檢測到了閃焰。然而，經過影像與顏色分析，我認為這些閃焰極可能來自未解析的低質量伴星，而非這些星體本身。;By being inspired by my prior researches on stellar flares, I developed a multi-algorithm machine learning approach aimed at significantly enhancing the efficiency of flare detection in light curves, thereby deepening our understanding of stellar flares from vast datasets.
I applied this approach to TESS 2-minute survey data from Sectors 1-72 to identify stellar flares.
Models trained with Deep Neural Network, Random Forest, and XGBoost algorithms, respectively, utilized four flare light curve characteristics as input features. Model performance was evaluated using accuracy, precision, recall, and F1-score metrics, all exceeding 94\%. Validation against previously reported TESS M dwarf flare identifications showed that my models successfully recovered over 92\% of the flares while detecting $\sim2,000$ more small events, thus extending the detection sensitivity of previous work. After processing 1.3 million light curves, the models identified nearly 18,000 flare stars and 250,000 flares. I present an extensive catalog documenting both flare and stellar properties. I found strong correlations in total flare energy and flare amplitude with color, in agreement with previous studies. Flare frequency distributions were analyzed, refining power-law slopes for flare behavior with the frequency uncertainties due to the detection incompleteness of low-amplitude events. I determined rotation periods for $\sim120,000$ stars thus yielding the relationship between rotation period and flare activity. I found that the transition in rotation period between the saturated and unsaturated regimes in flare energy coincides with the same transition in rotation period separating the saturated and unsaturated levels in coronal X-ray emission. Stellar X-ray emission increases more rapidly with flare luminosity in earlier-type and unsaturated stars, indicating more efficient coronal heating in these objects. Additionally, the models detected flares in white dwarfs and hot subdwarfs that are likely arising from unresolved low-mass companions. |
