Summary
In this chapter, we introduced a new recurrent neural unit: the LSTM unit. We showed how it is built and trained, and how it can be applied to a time series analysis problem, such as demand prediction.
As an example of a demand prediction problem, we tried to predict the average energy consumed by a cluster of users in the next hour, given the energy used in the previous 200 hours. We showed how to test in-sample and out-sample predictions and some numeric measures commonly used to quantify the prediction error. Demand prediction applied to energy consumption is just one of the many demand prediction use cases. The same approach learned here could be applied to predict the number of customers in a restaurant, the number of visitors to a web site, or the amount of a type of food required in a supermarket.
In this chapter, we also introduced a new loop in KNIME Analytics Platform, the recursive loop, and we mentioned a new visualization node, the Line Plot (Plotly) node...
