科研之友
微信
新浪微博
Facebook
分享链接
摘要
Northeastern Argentina is the forest area of greater importance in the country, concentrated in the provinces of Misiones and Corrientes, with Pinus spp. L., the species with higher production, which supplies raw materials to a large number of industrial activities. This highlights the need to implement forest management tools to make better decisions in investment and management of forests. Forest management models often use different techniques, including simulation, based on operational research, and econometric tools. Usually, the econometric techniques tend to be used for projections of prices and returns. An important class of models with longitudinal data is the family of Autoregressive moving average models, known as ARIMA, by its acronym in English, usually applied to describe trends and generate predictions from values passed from the series. In particular, the variation of prices of forest products is one of the main sources of uncertainty in forest planning. Nevertheless, the application of techniques and prediction models in the forestry area, especially at the South American region is still low. ARIMA Models exhibit good predictive short-term performance, although they lose ability to forecast in distant horizons and have some other disadvantages. Various autoregressive moving average models (ARIMA) based on Box-Jenkins methodology are proposed to predict future prices of four products for Pinus spp manufactured in Northeast Argentina. Estimations were carried out with time series of prices of the four products covering the period July 2002-September 2013. The proposed models predict future prices with forecast errors between 0,9% and 1,8%.
- 出版日期2014
