成分数据的组合预测

The Combination Forecast about Compositional Data

成分数据是一类具有复杂性质的数据,特别是它的预测研究在管理学与经济学中占有很重要的地位.组合预测则是近年来在预测中应用比较广泛的一种方法,它能够充分利用单预测模型的信息,提高预测精度,增强预测的稳定性,且具有较高的适应能力.本文首次把组合预测方法应用到成分数据的预测分析中,基于成分数据的一些基本性质,利用组合预测得到了较好的预测结果.

Compositional data is a kind of complex qualitative data, especially its prediction research plays an important role in management science and economics. The technology of combination forecast is widely used in the forecast, which makes full use of the single forecast model and makes progress on the prediction accuracy. In this paper, the combination forecast method is applied to the prediction of compositional data analysis, based on some basic properties of the compositional data. We can see from the example that using the combination forecast can get a better prediction result.