TY - GEN
T1 - RELATION BETWEEN FORECAST ERROR AND AGGREGATION LEVEL IN MARKOVIAN MODELS OF MOVEMENT.
AU - Rogerson, Peter
AU - Tromvoukis, Leonidas
PY - 1984
Y1 - 1984
N2 - This paper is concerned with evaluating the relative merits of alternative approaches to forecasting aggregate variables. 'Aggregate' may refer to variables that have been averaged over space, socioeconomic characteristics, or just about any other imaginable dimension. Authors focus on comparing the forecasts derived using two distinct approaches. In the first, a model is used at the disaggregate level and the aggregate forecasts are obtained by aggregating the disaggregate results in an appropriate manner. In the second approach, the same model is simply used at the aggregate level to derive forecasts.
AB - This paper is concerned with evaluating the relative merits of alternative approaches to forecasting aggregate variables. 'Aggregate' may refer to variables that have been averaged over space, socioeconomic characteristics, or just about any other imaginable dimension. Authors focus on comparing the forecasts derived using two distinct approaches. In the first, a model is used at the disaggregate level and the aggregate forecasts are obtained by aggregating the disaggregate results in an appropriate manner. In the second approach, the same model is simply used at the aggregate level to derive forecasts.
UR - https://www.scopus.com/pages/publications/0021662548
M3 - Conference contribution
AN - SCOPUS:0021662548
SN - 0876648308
T3 - Modeling and Simulation, Proceedings of the Annual Pittsburgh Conference
SP - 433
EP - 437
BT - Modeling and Simulation, Proceedings of the Annual Pittsburgh Conference
PB - ISA
ER -