|
| |
| Pål Børing, Arne Martin Fevolden and André Lynum |
| |
| ''Skills for the future – forecasting firm competitiveness using machine learning methods and employer–employee register data'' |
| ( 2021, Vol. 41 No.2 ) |
| |
| |
| This article investigates whether skills data can be used to forecast firm competitiveness. It makes use of an employer–employee register dataset consisting of detailed information about the educational background of all employees in the manufacturing sector in Norway and uses this data to predict the manufacturing firms' revenues five years into the future. The predictions are carried out by employing three machine learning models – lasso regression, random forest and gradient boosting. The results show that machine learning models using skills data can provide reasonably good forecasts of firm competitiveness. However, the results also show that these models become less reliable at the “extreme ends” and that they predicted extreme increases or decreases in revenues poorly. |
| |
| |
| Keywords: Lasso, Random Forest, Gradient Boosting, Skills, Education |
JEL: C5 - Econometric Modeling: General
L6 - Industry Studies: Manufacturing: General |
| |
| Manuscript Received : Dec 07 2020 | | Manuscript Accepted : Apr 09 2021 |
|
