Nonparametric Forecasting of the Manufacturing Output Growth with Firm-level Survey Data
Résumé
A large majority of summary indicators derived from the in- dividual responses to qualitative Business Tendency Survey questions (which are mostly three-modality questions) result from standard aggregation and quantification methods. This is typically the case for the indicators called balances of opin- ion, which are the most currently used in short term analysis and considered by forecasters as explanatory variables in lin- ear models. In the present paper, we discuss a new statistical approach to forecast the manufacturing growth from firm- survey responses. We base our predictions on a forecasting algorithm inspired by the random forest regression method, which is known to enjoy good prediction properties. Our al- gorithm exploits the heterogeneity of the survey responses, works fast, is robust to noise and allows the treatment of missing values. Starting from a real application on a French dataset related to the manufacturing sector, this procedure appears as a competitive method compared with traditional competing algorithms.
Origine : Fichiers produits par l'(les) auteur(s)