数据流聚类算法在数据处理中的应用

基于监测数据进行分析,采用特殊投影和拟合的数据流聚类算法,该算法结合拟用算法和滑动窗口技术进行聚类预处理,掌握数据流的变化趋势。此外,通过桥梁健康监测仿真实验,得出拟合数据具有更高的精确性,分段拟合的数据R2接近1,m为0.890 4＜0.5,解决了高维数据流的降维问题,缩短了数据处理时间,且提高效率39.1以上,占用空间仅为6.7%。

Feature-based Data Stream Clustering Algorithm Projection and Fitting in Data Processing

The trend of data stream is analyzed based on monitoring data using a special projection clustering and fitting algorithm,which combines the pre-processing algorithms and sliding window technique.The bridge health monitoring simulation experiments are performed with fitting data of higher accuracy (segment R2 at distance close to 1,m=0.890 4＜0.5),thus solving the dimensionality reduction of high dimensional data stream,and reducing the data processing time with an increase in efficiency by 39.1% and a space occupation of only 6.7%.