状态扩展元胞自动机模型在时空数据挖掘中的应用

Application of Extended State Cellular Automata to Spatiotemporal Data Mining

引入状态扩展元胞自动机模型对时空数据进行挖掘,其核心是引入可以量化的属性和不可量化的状态对元胞状态进行扩展,解决时空数据挖掘中数据稀疏性和属性数据交互性问题,采用遗传算法寻找元胞自动机模型的最优规则。实验结果表明,对于复杂的非线性和数据稀疏性问题,利用该方法能得到比传统方法更好的结果。

The paper introduces an extended state cellular automata （CA） model to spatiotemporal data mining（STDM）. The core of the model adds numerable and uncountable attribute to the cell and intends to resolve the problem of the sparse data and large attribute in- formation interaction in the spatial and spatiotemporal data mining tasks. The preliminary experiment shows the approach is suited for the nonlinear problems, even in the face of sparse data. They can tackle problems of previously prohibitive complexity and also improve previous approaches. The paper advises the method in combination with domain knowledge and other data mining techniques offer a chance to discover nonlinear spatiotemporal relationships.