Undiagnosed samples aided rough set feature selection for medical data

摘要: Medical data often consists of a large number disease markers. For medical analysis, some markers are not helpful and sometimes even have negative effects. Therefore, applying feature selection is necessary as it can remove those unimportant Among many methods, rough set based (RSFS) has been widely used. Unlike other RSFS completely data-driven. It does require any information like probability distributions. Traditional methods extract the only from diagnosed samples. they usually samples to achieve good performance. However, in real applications, limited, yet undiagnosed large. Motivated by semi-supervised learning methodology, this paper, we propose novel method which learn both This called aided (USA-RSFS). Its main benefit reduce requirement on help ones. Finally, promising performance USA-RSFS validated through experiments datasets.