
Forecaster Gas
Provide daily withdrawal estimates and monthly planning for points of delivery
Published by
Accenture

Forecaster Gas
Published by
Accenture
Features
Statistical methods
The application offers a full range of statistical methods, including highly, market-specific forecasting methods such as authority profiles and non-daily metered algorithms.
Produce an accurate forecast (short, medium and long term), to reduce physical unbalancing and related costs
Minimize the statistical know-how required to achieve accurate forecast
Import historical time series from TSO and 3rd party weather condition forecasting
Industrialize the process, allow business users to have strong control over it, reduce effort to focus on results and business decisions
Take into account Energy Market peculiarities (uses different version of TSO's allocation, NDM meters, direct customers, market detail under the DP, standard profiles, contract volumes, other TSO information can be used as drivers)
Advanced Analytics
Forecast individual time series: classical time series models, regressions, autoregression, neural networks, GAM models, free syntax models
Cluster points/meters in groups with similar behavior using hierarchical clustering algorithms
Access library of pre-configured models for the most common behaviors of gas demand
Select most effective model in library with back testing simulation logic
Statistical methods
The application offers a full range of statistical methods, including highly, market-specific forecasting methods such as authority profiles and non-daily metered algorithms.