Normalize helps us scale data in a DataSet from any input range to a normal output range. Typically, a normal range is between 0 and 1. However, these boundaries can be modified with the min and max parameters. A concise example of normalisation is shown below by applying it to a DataSet containing five points. The input data is shown on the left, and the normalised result is shown on the right.

DataSet

Identifier
Data
1
1.00
2
2.00
3
3.00
4
4.00
5
5.00

DataSet

Identifier
Data
1
0.00
2
0.25
3
0.50
4
0.75
5
1.00

When there are multple columns of data, normalisation is applied per column. This preserves the relationship of data within a column across all points, but will alter the shape of the data across the point. In the example below, we might imagine the first column is Loudness in dB, and the second column is Pitch in hertz. Observe the differences in range between the first and second column after normalisation.

DataSet

Identifier
Data
1
-12.00
220.00
2
-24.00
440.00
3
-36.00
880.00
4
-48.00
1760.00
5
-60.00
3520.00

DataSet

Identifier
Data
1
1.00
0.00
2
0.75
0.07
3
0.50
0.20
4
0.25
0.47
5
0.00
1.00
Last modified: Tue Aug 23 14 by James Bradbury
Edit File on GitHub