关键词: 时间序列数据挖掘, 地铁门电机数据, 模式识别, 主成分分析, 拓展符号聚集近似(ESAX)

Abstract: Aiming at the problem that the sub-health status of train door is difficult to identify, a sub-health status identification method based on time series data mining was proposed. First of all, the angle, speed and current data of the door motor were discretized by combining the multi-scale sliding window method and Extension of Symbolic Aggregate Approximation (ESAX) algorithm. And then, the features were obtained by calculating the distance with the template under the normal status in of train door, in which the Principal Component Analysis (PCA) was adopted to reduce feature dimensions. Finally, combining with basic features, the hierarchical pattern recognition model was proposed to identify the sub-health status from coarse to fine. The train door real test data were taken as example to verify the effectiveness of the proposed method. The experimental results show that the proposed algorithm can distinguish sub-health state effectively, its recognition rate can reach 99%.

Key words: time series data mining, engine parameter of train door, pattern recognition, Principal Component Analysis (PCA), Extension of Symbolic Aggregate Approximation (ESAX)