Definition
1. Data preprocessing is a data mining technique that involves transforming raw data into an understandable format. Real-world data is often incomplete, inconsistent, and/or lacking in certain behaviors or trends, and is likely to contain many errors. Data preprocessing is a proven method of resolving such issues. Data preprocessing prepares raw data for further processing.
2. Data preprocessing describes any type of processing performed on raw data to prepare it for another processing procedure. Commonly used as a preliminary data mining practice, data preprocessing transforms the data into a format that will be more easily and effectively processed for the purpose of the user -- for example, in a neural network. There are a number of different tools and methods used for preprocessing, including: sampling, which selects a representative subset from a large population of data; transformation, which manipulates raw data to produce a single input; denoising, which removes noise from data; normalization, which organizes data for more efficient access; and feature extraction, which pulls out specified data that is significant in some particular context.
Abbreviation
Synonyms
Superterms
Data processing
Subterms
Sources
https://www.techopedia.com/definition/14650/data-preprocessing (1.); http://searchsqlserver.techtarget.com/definition/data-preprocessing (2.)
Author: Simon Waterstradt
