Classification – A General theory and Efficiency in Medical Diabetes Dataset

Mining is the process of extracting efficient and correct information from a large sample of metadata. Data mining indicated gathering data from preexisting data sources. Classification is a branch of data mining method that is based on the supervised learning mechanism of categorizing data. This paper proceeds with the supervised learning algorithms that can exclusively deal with categorical and multi-variated data. The classification algorithms Linear Regression of Statically testing, Random Forest of Decision Tree and Rule Generation Decision Table are tested in 10 cross-fold validation to choose the best performing algorithm. It mainly focuses on the algorithm’s tenacity in handling outliers in the medical field using the diabetes dataset. The Rule Generation-based Decision Table algorithm performs well with the multi-variate and categorical diabetes dataset with an accuracy of 81.1% in result calculation.