Research on the Technology of Forest Volume Prediction

The self-development system of forest growing-stock prediction is based on VB p latform. It uses GM (1, 1)model, compound interest formula and BP artifical neural network model to make macroscop ic forecasting of forestgrowing-stock respectively on the basis of management inventory data of forest resources of Sanming, FujianProvince. The forecasting result of three methods showed that BP artificial neural network model is fitted the best,followed by the Grey model, and the compound interest formula had the largest average relative error . Finally thispaper analyzes the p ros and cons of the three methods and exp lores further op timization method.