关键词: 特征, 聚类和分类, 时间序列, 油田测试优化

Abstract: Note the oilfield production cross section in the dynamic test in the well election does not have a suitable reference standard issue, the paper proposes a well production division method in accordance with the feature-based clustering and the classification. First, the method selects and processes the well production time series data. And then the characteristic sets are extracted for clustering from time series data. In the end the clustering results develop the oilfield test optimization. Experimental result shows that the characteristics extracted from time series can be a good representation of the fluctuations in oil production, obtain better classification results and be important for guiding the oilfield test.

Key words: characteristics, clustering and classification, time series, the oilfield test optimization