Manual: 3.2. Prepare Your Data

Machine learning works by analyzing historical data to find some pattern and to store that pattern in a mathematical model. For this to work, the methods require both the historical data itself and some extra information about the data. This section will describe how to prepare all this so that the learning can take place.

The situation that we want to model will generally have many properties that are measured by sensors and that we can include in the model. The first step is to choose which of the available sources of data will be included in the analysis. As a second step, each of these data sources needs to be characterized with some essential information such as its allowed range of values. The third and last step is to prepare a table of historical values for each of the data sources.

With all of this data in place, the analysis and learning can begin.