Improving remote sensing crop classification by argumentation-based conflict resolution in ensemble learning

作者：

Highlights：

• A hybrid system for crop classification from satellite images is proposed.

• A novel method for conflict resolution in ensemble learning is developed.

• Argumentation technology performs dialectical analysis on debatable instances.

• Agricultural expert knowledge is merged with rules extracting from base learners.

• Classification accuracy and transparency of the decision are increased.

摘要

•A hybrid system for crop classification from satellite images is proposed.•A novel method for conflict resolution in ensemble learning is developed.•Argumentation technology performs dialectical analysis on debatable instances.•Agricultural expert knowledge is merged with rules extracting from base learners.•Classification accuracy and transparency of the decision are increased.

论文关键词：Crop classification,Ensemble learning,Defeasible argumentation,Agricultural expert knowledge,Rule extraction

论文评审过程：Received 26 March 2016, Revised 16 June 2016, Accepted 25 July 2016, Available online 28 July 2016, Version of Record 3 August 2016.

论文官网地址：https://doi.org/10.1016/j.eswa.2016.07.037