|
|
The forecast module (ic-fc) is conceived to provide a clear characterization of the most common forecasting data. During the ontology requirement specification process with the InterConnect partners, we noticed that a distinction was needed between point forecasts versus stochastic forecasts, as well as the various ways to express stochastic forecasts. Therefore, the ic-fc module takes into account this distinction. It reuses the ic-data module (see Section 4.2), which defines time-series and data-points that are important elements of forecasting. It also reuses the ic-topology module (see Section 4.9), which defines the forecast location (geographical but also topological, for example, the grid segment). Figure 7 shows the concepts that we designed for defining forecasts. In order to keep the figure readable, we focused on the new concepts and we did not extensively visualize the concepts already presented in Figure 6.
|
|
|
|
|
|
![Forecast.drawio](uploads/0a573923454b8199b0f0589430819425/Forecast.drawio.png)
|
|
|
|
|
|
***Forecast types:***
|
|
|
- ic-fc:StochasticForecast is the base type for all forecasts that have stochastic or probabilistic data points. This means we place restrictions on the type of data points a stochastic forecast consists of. We have three predefined stochastic forecasts.
|
|
|
- ic-fc:GaussianStochasticForecast contains a forecast following the Gaussian distribution. All its data points are therefore of the type GaussianDataPoint as defined in the ic-data module. Each data point of this forecast type therefore has the mandatory ic-fc:hasStandardDeviation property.
|
|
|
- ic-fc:QuantileForecast gives the option of manually defining the quantile for which a particular value is intended. The respective quantiles can be added via the ic-fc:hasQuantile property.
|
|
|
- if-fc:TrajectoriesForecast contains various simple time-series that describe possible alternatives. Each individual time-series is an instance of the ic-data:TimeSeries class.
|
|
|
- if-fc:PointForecast contains exclusively simple data points without a stochastic or probabilistic element. Each data point is expressed as an instance of ic-data:DataPoint.
|
|
|
|
|
|
***Forecast properties:***
|
|
|
- ic-fc:hasQuantile, which assigns to the data point the percentage of values that are below this value. In other words, a data point with quantile 90 indicates that 90% of other measurements are (estimated to be) lower.
|
|
|
- ic-fc:hasStandardDeviation is a mandatory property for Gaussian forecast data points. The standard deviation (i.e., the square root of the average of the squared deviations of the values subtracted from their average value) can be described with this property.
|
|
|
|
|
|
|
|
|
|
|
|
|