Intelligent Forecast for Cucumber Fusarium Wilt Combining Case-based Reasoning With Self-organizing Maps

Combining self-organizing maps (SOM) with case-based reasoning (CBR), a hybrid intelligent forecast method for CFW (cucumber fusarium wilt) is presented. Different from the traditional similar case retrieval, this method performs case classification with trained SOM network and then figures out a similar case set using a proposed case similarity metric. A classification accuracy of 97.22% was achieved by the integrated SOM network in the classification performance test. From CFW forecast experiments, the optimal interval of dissimilarity threshold R for this method is inferred. Comprehensive analysis shows that this hybrid forecast method can effectively provide reliable reasoning data for CFW forecast and assist decision-making of CFW prevention and treatment measures.