July 2017
Beginner to intermediate
378 pages
10h 26m
English
IoT data is commonly captured at regular intervals, and the position of the value in the series can have a meaning. A dataset captured at regular intervals is called a time series. Time series present their own complications for ML.
As mentioned, the position of the value in the series can have meaning beyond just the value itself. Saying something has meaning in ML is basically saying it has predictive value. It should be captured as a feature in the model input so that a model can incorporate it into the learning process.
This can be done by transforming the data so that meaningful elements of the time series are captured as additional features in the dataset. The following are some examples:
Read now
Unlock full access