Community Research and Development Information Service - CORDIS

Abstract

Multilayer perceptron neural networks are used to implement forecast models. The quantity to be forecast is annual industrial production, for each Member State of the European Union and some 150 industries. Predictors are selected amongst many leading indicators available. For any given industry, a forecast model which is common to every Member State was searched. If a couple of easily checkable data-driven conditions are fulfilled, multilayer perceptron models have better average performance than the method currently used at the Statistical Office of the European Communities (EUROSTAT).

Additional information

Authors: ALVAREZ M, JRC Ispra (IT);AUGER J-M, JRC Ispra (IT);VARFIS A, JRC Ispra (IT)
Bibliographic Reference: Article: Research in Official Statistics (1996)
Record Number: 199710173 / Last updated on: 1997-04-01
Category: PUBLICATION
Original language: en
Available languages: en