A forecasting engine arranged to generate a forecast for a set of historic time-series data. The engine includes one or more one or more generically structured core algorithmic components providing a core function in a forecasting heuristic algorithm, and one or more generically structured optional algorithmic components providing an optional function in the forecasting heuristic algorithm. Each algorithmic component is individually tuned in a predetermined sequence, and the first algorithmic component in the sequence performs a tuning process on the set of historic time-series data. Subsequently, algorithmic components are tuned using time-series data conditioned by ail of the tuning processes previously performed in the predetermined sequence. The entire sequence of algorithmic components is arranged to collectively provide conditioned data which is used to generate a forecast.

 
Web www.patentalert.com

< Context-based failure reporting for a constraint satisfaction problem

> Decision support systems and methods

> Automated web-based targeted advertising with quotas

~ 00532