Model Improvements
AI Model Collection
We have gathered all our new AI/Machine Learning models under AI Model Collection (AMC). Right now we have two AI-ML-models available for each enery source, to optimize your forecasts.
AI Model Collection is a subscription of our existing AI/ML models as well as future models and improvements of these. Our AI/ML development team are continuously searching for new- and improvements in existing models.
MOS – Model Output Statistics
Improves weather forecast input by using Model Output Statistics related to weather observations. Improvements by up to 7.3% with MOS.
Model Output Statistics for wind speed » Model Output Statistics for temperature »
Post-processing
Calculates an error estimate by observing historical biases for each time step which is used as a corresponding correction for future forecasts. Improves the forecast error for day ahead for electricity demand load forecast by up to 9.5%.
Post-processing of load forecasts »
Software Features
Market Data
Quality Service
Aiolos Quality Service is our advanced warning system that can create warning messages to the user if there is something wrong with the forecast series, such as bad or missing in-data/weather forecasts etc.
Quality Service enables you to filter imported and exported data to Aiolos so that you base your forecasts on correct data. If data (also the forecasts) is missing or are of low quality, the system will send you warnings about this so you can find and fix the problem.
Vitec Secure Hosting
Vitec hosts thousands of clients in their everyday operations. When we operate and manage Aiolos Forecast Studio, you can focus on your core business. The service also includes daily checks of imported and exported data to Aiolos, for example weather data. We also check that the weather data used are of good quality.
Advanced Realtime Corrections
HOME Live
Weather Point Optimizer
Weather Point Optimizer automatically finds the best correlation between your forecast series and all available weather points. We are also optimizing which weather suppliers to use for each forecast series. To describe it shortly, the module is optimizing how to handle weather data for your forecast series.
Data Improvements
Deutscher Wetterdienst
Number 9 in the included weather forecast suite.
Deutche Wetterdienst is the latest weather forecast institute to be brought into Aiolos. All institutes are tested by our model development department to ensure that they bring value (read: lower the MAE) to you as a customer.
Foreca
Number 8 in the included weather forecast suite.
Foreca has over 20 years of experience and know-how in weather forecasts. Their customers include some of the world's largest and most advanced media companies.
Get help from our team of experts
DayType Calendar Tuning
By using the DayType Calendar in an optimal way for each forecast series, there are a lot of improvements in terms of forecasting accuracy to be made. One example on what to optimize is the usage of weekends vs individual Saturdays/Sundays and how different settings of these day-types affect the forecast accuracy. Contact us to get more info on what we can help you with.
Aiolos Quality Service Tuning
Aiolos Quality Service is the module that filter all incoming data, all internal processes during the making of the forecast and all outgoing data. This is of huge importance to assure that the models have the best possible accuracy. Vitec staff can help with tuning of these filters to find a suitable level for your company.
Full Forecasting Service
Today we are helping some of our customers with the whole process of making an accurate forecast. The outcome of this service is an improved forecast optimized by us, so you can spend your time on other things. We can offer a full-service or a part-time service where we for example check settings/data two hours per week.
