Data based segmentation and summarization for sensor data in semiconductor manufacturing
作者：
Highlights：
• Data were classified to discrete-valued or continuous-valued groups.
• Discrete-valued data were segmented with abnormal difference.
• Continuous-valued data were segmented with free knot spline with knot removal.
• Segmentation and summarization were performed based on linearity and parsimony.
• Segmentation had 38.54% error decrease compared to the conventional method.
摘要
•Data were classified to discrete-valued or continuous-valued groups.•Discrete-valued data were segmented with abnormal difference.•Continuous-valued data were segmented with free knot spline with knot removal.•Segmentation and summarization were performed based on linearity and parsimony.•Segmentation had 38.54% error decrease compared to the conventional method.
论文关键词：Semiconductors,Sensor data,Segmentation,Summarization,Free knot spline with knot removal,Time series sensor segmentation
论文评审过程：Available online 15 November 2013.
论文官网地址：https://doi.org/10.1016/j.eswa.2013.11.001
