Agricultural Data Classification Based on Rough Set and Decision Tree Ensemble

作者：Shi, Lei; Ma, Xinming^*; Duan, Qiguo; Weng, Mei; Qiao, Hongbo

来源：Sensor Letters, 2012, 10(1-2): 271-278.

DOI：10.1166/sl.2012.1857

• Rough Set
• Decision Tree
• Agricultural Data

Because of explosive growth of agricultural data available, the classification of agricultural data is becoming one of the most important topics in the field of precision agriculture technologies. Rough set theory is a formal mathematical tool to deal with imprecision, vagueness and uncertainty. Ensemble learning is one of widely used techniques for improving the generalization ability of classification models. In this paper, rough set and decision tree ensemble is combined to construct the hybrid classification approach to increase the performance of the prediction of agricultural data. In the proposed approach, rough set theory is employed as a preprocessor to reduce the redundant attributes with no information loss, then a collection of decision tree classifiers is built from bootstrap samples to construct ensemble. An experimental evaluation of different methods is carried out on several publicly available agricultural datasets. The experimental results indicate that the proposed method achieves significant performance improvement.

• 出版日期2012-2
• 单位河南农业大学