LIONcommunity usage case
LIONbook: Learning requires a method.
This is an exercise associated with the LIONbook
and with * Tutorial video #1
You may want to watch the video before starting.
The data are about assessing the heating load and cooling load requirements of buildings
(that is, energy efficiency )
as a function of building parameters. The real experiment could be:
i) Collect data about houses
(geometry, materials, etc.), ii)
read their energy meters to prepare a table with training data.
iii) Train a model.
The model can be used to:
- Predict the energy consumption for a new house project.
- Find the optimal design of a building to minimize energy consumption.
Using the K Nearest Neighbors model in LIONoso is seamless if you are already familiar
with the user interface. Do not worry if you do not know the math behind KNN, you can still get its value.
Be sure to read the
first pages in the LIONoso manual to familiarize yourself with the user interface (the manual is included
in the software distribution). Click on the images to get a larger version.
Step 1: Load data files and KNN, split data into training and testing files
To load data file, drag the "CSV file" icon to the workbench to the right,
drag the "Split table" node and connect them by drawing an arrow (or drop one on top of the other one).
Try with the file energy-efficiency.csv.
You will get a random split of the initial file into a training file and a separate testing file .
To load the KNN factory (the creator of KNN models) drag the
"KNN factory" to the workbench to the right, and connect it to the training file.
Click "Start training" to create the KNN model (same icon but no gear symbol).
Step 3: Visualize predicted outputs on testing set
Connect the table with the testing data to the KNN node to get the predicted output values.
Note: if the standard name of the component is getting large, you can pick your favorite name ("test" in the example).
Right-click the icon with the prediction table and select "New panel - Bubble". Click on Dashboard to visualize the
Bubble-chart. Drag the variables HeatingLoad-Target (the target value given in the table) and HeatingLoad (the predicted value)
if not already visualized.
Step 4: Analyze test data
Connect the error analyzer to the test data for additional analysis (root mean square error, mean absolute error)