Survey on Data Preprocessing Concept Applicable in Data Mining

Mathew Ngwae Maingi

Abstract: Real world data is highly prone to outliers commonly known as data noise. This occurrence usually causes a problem of missing values or maybe data full of inconsistencies thus resulting to a poor quality data. Poor quality data is unreliable and fake since it never upholds data integrity issues. Principally, computer users wish to harvest data that is reliable and of high integrity and thats where the concept of data preprocessing comes in since quality decisions are directly proportional to quality data. Data preprocessing deals with data preparation and data transformation, and seeks to improve the overall process of data mining and at the same time make the process of knowledge discovery more efficient. This paper therefore focuses on surveying different data preprocessing techniques as used in data mining, exhaustively outlining their major purposes in knowledge discovery process.

Keywords: Data, noise, integrity, preprocessing, transformation