Improved Algorithm for Case Feature Weight Based on Sensitivity Analysis
Shaowen Li · 2010
Method for weight of case-based reasoning feature was studied.A method based on BP neural network and sensitivity analysis was put forward,and the algorithm of the weight for the feature in case was improved.BP neural network about feature and decision was constructed,after training and learning,the input nodes were deleted one by one,the output of sensitivity was analysed and weight was calculated.The case on Red-seed Watermelon was tested,which shows the algorithm is effective.