Feature selection for classification: A review

Abstract

However, these collected data are usually associated with a high level of noise. There are many reasons causing noise in these data, among which imperfection in the technologies that collected the data and the source of the data itself are two major reasons. For example, in the medical images domain, any deficiency in the imaging device will be reflected as noise for the later process. This kind of noise is caused by the device itself. The development of social media changes the role of online users from traditional content consumers to both content creators and consumers. The quality of social media data varies from excellent data to spam or abuse content by nature. Meanwhile, social media data are usually informallywritten and suffers from grammaticalmistakes, misspelling, and improper punctuation. Undoubtedly, extracting useful knowledge and patterns from such huge and noisy data is a challenging task.