According to the feature of market fluctuations in the price of sugar, an optimization algorithm based on improved genetic neural network training was proposed in this paper. A population optimization model on adaptive crossover and mutation operator and niche was designed, by applying gray theory and technology, the sugar price data was processed. A multi-dimensional learning sample and teacher sample for improved genetic neural network training was constructed. Finally, the trend of sugar prices of 1-2 weeks in year 2008 to 2009 was predicted by cases, the comparison of the forecast algorithm versus gray linear systems, S-BP, SGA-BP algorithm showed the integrated optimization of forecast accuracy and forecast effect.
展开▼
