DATA CLEANING FOR DATA MINING AND APPLICATIONS ON TURKISH CLASSICAL MUSIC DATA

As a result of increasing data sets, keeping data under control and analyzing them have become compulsory. It is a necessity to convert the raw data into knowledge with data mining. Managing knowledge adds value to data. Furthermore, it is a basis for innovation. Even though data science is generally used in production, economics, informatics, health etc., it is possible to implement it in all sectors where data exist. In this study, a statistical analysis is conducted on a data set of Turkish Music repertoire which was downloaded with a Java script programme. This data set consists of 43936 data. For the statistical analysis the topic is selected in rhythmic patterns which is the structure of music called “us？l” in Turkish Classical Music. This study concerns with the relationship between “us？l” and prosodic patterns called “ar？z” and the importance of data cleaning in data mining process. The result of the statistical analysis shows that composers who lived in the early 20th century used “ar？z” more than others