Presentation #108.07 in the session “Missions and Instruments (Poster)”.
Machine learning techniques are useful tools for sifting through large astronomical datasets. We apply a principal component analysis (PCA) and an unsupervised Random Forest (uRF) to the second release of the Chandra Source Catalog (CSC 2) to search for unusual X-ray sources. The CSC2 contains ~317,000 X-ray sources observed by NASA’s Chandra X-ray Observatory (CXO) through 2014. We limit our analysis to high-significance (detection significance ≥ 7.5) sources to ensure that existing observations can offer a “first look” follow-up. We found 119 sources that were consistently identified as outlier sources by the uRF across 100 applications of the algorithm, many of which are well-observed but have not been individually investigated.