Orchestration of Environmental Services
The objective is to develop techniques for the orchestration of services, including (i) matchmaking between the environmental service nodes researched by the search engine, with the goal to find the nodes that provide the most adequate and most reliable information needed to handle the requests of the decision support and user interaction techniques; (ii) connecting the selected nodes to establish the data, knowledge and control flow.
To be able to select the most reliable and most exact environmental data providers among those that offer their data on the web or provide the data via web services, confidence and uncertainty metrics are about to be developed. The confidence metric estimates the trustworthyness of the provider, based on external evlauation of the correctness of its data and its metadata. The uncertainty (or impresion) metric estimates the accuracy of the data, which may be affected by faulty measurements, data gaps, data transmission noise, error range of the forecast, etc.:
Experts can intervene to adjust the outcome of the metrics. They are supported by time-tested tools for inclusion of forecasters’ knowledge and expertise into numerical weather forecast models:
The work in the second year of PESCaDO focuses on ensemble-based chemical weather forecasting models for the optimal merge of the data from all sources encountered in the web. The data from each source will receive a dynamic weight calculated in accordance with a performance coefficient metric using machine learning techniques.
Ari Karppinen <email@example.com>