From the course: Applied AI for IT Operations (AIOps)
Join today to access over 22,600 courses taught by industry experts or purchase this course individually.
Forecasting future service loads with Keras - Python Tutorial
From the course: Applied AI for IT Operations (AIOps)
Forecasting future service loads with Keras
- [Instructor] We predicted for the test dataset in the previous video. In this video, we will forecast future service loads for a full week using the model we built in the earlier videos. One important factor to consider here is this. In the case of the test data we used, each data point we wanted to predict, we had 168 lookback data points from the actual data. But if we want to predict in the future, we can only predict the first point, since only that point has a complete lookback from actual data. If we reduce our lookback size, it will impact the overall accuracy of our predictions. How do we predict data points beyond the first point? To do this, we iteratively predict each data point and then add that predicted point into the lookback dataset for the next point. For example, to predict a new forecast point P1, we use the last 168 points from the test dataset. Then to predict the next point, which is P2, we use a lookback that contains P1 and the last 167 points from the test…
Practice while you learn with exercise files
Download the files the instructor uses to teach the course. Follow along and learn by watching, listening and practicing.
Download courses and learn on the go
Watch courses on your mobile device without an internet connection. Download courses using your iOS or Android LinkedIn Learning app.
Contents
-
-
-
-
-
-
Time series forecasting2m 27s
-
(Locked)
Recurrent neural network (RNN) and long short-term memory (LSTM)1m 37s
-
(Locked)
Preparing sequence data4m 8s
-
(Locked)
Building an LSTM model with Keras2m 19s
-
(Locked)
Testing the time series model2m 30s
-
(Locked)
Forecasting future service loads with Keras3m 10s
-
-
-
-
-