{{merge to|Canonical form|discuss=Talk:Canonical form#Proposed merge with Data normalization|date=November 2015}}
{{more|Normalization#Technology{{!}}sub-list: Normalization (technology) |other forms of data}}
'''Data normalization''' is the process of reducing data to its [[canonical form]].
For instance, [[Database normalization]] is the process of organizing the [[Field (computer science)|fields]] and [[Table (database)|table]]s of a [[relational database]] to minimize [[Data redundancy|redundancy]] and dependency. In the field of [[software security]], a common [[Vulnerability (computing)|vulnerability]] is unchecked [[malicious input]]. The mitigation for this problem is proper [[input validation]]. Before input validation may be performed, the input must be normalized, i.e., eliminating encoding (for instance [[Character encodings in HTML|HTML encoding]]) and reducing the input data to a single common [[character set]].
Other forms of data, typically associated with [[signal processing]] (including [[Audio signal processing|audio]] and [[Image processing|imaging]]), can be normalized in order to provide a limited range of values within a norm.
From a classifier point of view, a very popular way to normalize data consists on scaling input vectors into a 0 to 1 scale.
