Preprocessing

Raw data collected from a data source usually presents different particularities, such as data range, sampling, and category. Some variables result from measurements while others are summarized or even calculated. Preprocessing means to adapt these variable values to a range that neural networks can handle properly.

Regarding weather variables, let's take a look at their range, sampling, and type:

Variable

Unit

Range

Sampling

Type

Mean temperature

º C

10.86 – 29.25

Hourly

Average of hourly measurements

Precipitation

mm

0 – 161.20

Daily

Accumulation of daily rain

Insolation

hours

0 – 10.40

Daily

Count of hours receiving sun radiation

Mean humidity

%

45.00 – 96.00

Hourly

Average of hourly measurements

Mean wind ...

Get Deep Learning: Practical Neural Networks with Java now with the O’Reilly learning platform.

O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.