Itron Inc. has released MetrixND 4.2, an energy forecasting tool. The new, enhanced features of the tool include additional model building functions, more graphic capabilities that can be customized by the user, and additional statistics that assist in the evaluation of model performance.
MetrixND can be used for many forecasting processes, including short-term and long-term energy and demand forecasting, price forecasting, individual customer load forecasting and financial forecasting. First released in 1997, Itron’s MetrixND has more than 700 users from 160 utilities and energy companies in nine countries.
Designed to take advantage of advanced Windows capabilities, MetrixND offers an intuitive user interface while drag-and-drop architecture streamlines the development of forecasting variables and models. MetrixND presents a range of modeling techniques including regression and neural networks. Evaluation graphs, diagnostic statistics and reports are readily available to assist in developing and analyzing forecasts. The interactive design strategy enables forecast analysts to quickly evaluate alternative models and select the model that works best for the data provided.
“We've used MetrixND since 2002 and I've discovered it is one of the best tools to create accurate forecasting models," said Lori Cifuentes, manager, load research & forecasting for Tampa Electric Co. in Florida. “The enhanced model building functionality allows us to develop more dynamic and integrated models at monthly, daily and hourly levels. MetrixND is very easy to use and the reporting and graphing capabilities save us time.”
MetrixND is at the core of Itron’s automated forecasting solutions. These systems generate the forecasts used to schedule energy delivery and procurement in real-time markets and provide the foundation for financial forecasts. “The strength of MetrixND lies in its ability to produce accurate forecasts for a full-range of energy forecasting problems, from short-term power system operations to long-term financial and facilities planning. If it can be measured, then MetrixND can be used to forecast future values,” adds Monforte.