A local weighting method to the integration of neural network and case based reasoning

Our aim is to build an integrated learning framework of neural network and case based reasoning. The main idea is that feature weights for case based reasoning can be evaluated using neural networks. In our previous method, we derived the feature weight set from the trained neural network and the training data so that the feature weight is constant for all queries. In this paper, we propose a local feature weighting method using a neural network. The neural network guides the case based reasoning by providing case-specific weights to the learning process. We developed a learning process to get the local weights using the neural network and showed the performance of our learning system using the sinusoidal dataset.