Comparative Study of Outlier Detection Algorithms via Fundamental Analysis Variables: An Application on Firms Listed in Borsa Istanbul

摘要: In a data set, an outlier refers to point that is considerably different from the others. Detecting outliers provides useful application-specific insights and leads choosing right prediction models. Outlier detection (also known as anomaly or novelty detection) has been studied in statistics machine learning for long time. It essential preprocessing step of mining process. this study, process applied identifying top 20 firms. Three algorithms are utilized using fundamental analysis variables firms listed Borsa Istanbul 2011-2014 period. The results each algorithm presented compared. Findings show 15 identified by three methods. KCHOL SAHOL have greatest number appearances with 12 observations among these By investigating results, it concluded makes firm lists due differences their approaches detection.