基于分段线性方法的瓦斯浓度时间序列模式表示

Pattern Representation of Gas Concentration Time Series Based on Piecewise Linear Method

  • 摘要: 指出直接采用原始瓦斯浓度时间序列进行短期浓度预测、相似性查询、时间序列分类和聚类等数据挖掘工作不但效率低下,而且会影响时间序列数据挖掘的准确性和可靠性;提出了一种采用分段线性方法的时间序列模式表示方法。采用分段线性表示方法对瓦斯浓度时间序列进行模式表示后可换来较小的存储和计算代价,只保留了时间序列的主要形态,去除了细节干扰,更能反映出时间序列的自身特征,有利于提高数据挖掘的效率和准确性。

     

    Abstract: The paper indicated it not only has low efficiency of data mining to directly use original gas concentration time series to forecast short-term concentration,query similarity and classify and cluster time series,but also affects accuracy and reliability of data mining of time series.It put forward a pattern representation method of gas concentration time series based on piecewise linear method.By use of the piecewise linear method to make pattern representation of gas concentration time series,gas data storage and cost of count are smaller,only main form of gas concentration time series is reserved,and detailed interferes are went out,so the results of pattern nopresentation can reflect itself characteristics of gas concentration time series and and the method is good for improving efficiency and accuracy of data mining.

     

/

返回文章
返回