提出一种改进的混合压缩相符预测器.混合压缩相符预测器采用两阶段压缩过程:先将部分序列样本压缩成模型以知识的形式保存;再将知识传递给后续样本用于置信预测.混合压缩相符预测器不仅能提高计算效率,还提供巧妙的邻近性度量,从而极大地提高了第二阶段的预测效率.以田纳西-伊斯曼化工过程为例,验证了该方法的有效性和高效性. An improved hybrid compression match predictor is proposed.The hybrid compression match predictor uses a two-stage compression process: part of the sequence is compressed into a model and stored in the form of knowledge, and then the knowledge is passed to the subsequent samples for confidence prediction. The coincidence predictor not only can improve computational efficiency, but also provide ingenious proximity measure, which greatly improves the prediction efficiency of the second stage.With the Tennessee - Eastman chemical process as an example, the validity and efficiency of the method is verified .