基于数据立方体的评估特征项生成方法

Generation Method of Evaluation Features Based on Data Cube

针对传统评估指标生成模式无法适应作战态势快速变化的问题,提出一种基于数据立方体的评估特征项生成方法。以兵棋推演数据为基础,基于数据立方体框架构建评估特征项生成模型,利用不同维度组合生成评估特征项。围绕数据立方体中存在的“维度爆炸”问题,利用维度组合裁剪模型来缩小搜索空间,通过特征项标识算法进行唯一标识,并采用移动时间窗口提取评估特征数据。实验证明,该方法生成的评估特征项能够有效支撑对特定问题的评估。

For the problem of traditional generation pattern of evaluation indicators not adapting to rapid change of combat situation, a generation method of evaluation features based on data cube is proposed. Based on wargaming data, generation model of evaluation features is built in the data cube framework, and evaluation features are generated by different dimensions combination. Around the problem of dimension explosion in data cube, dimension combination reduction model is used to narrow the search space, and unique identification is calculated by feature identification algorithm, and evaluation feature data is extracted by moving time window. The test result show that evaluation features generated by the method can support special problem evaluation effectively.