Feature-Weighted CBR with neural network for symbolic features

Authors:
Sang Chan Park;Jun Woo Kim;Kwang Hyuk Im
Affiliations:
Korea Advanced Institute of Science and Technology (KAIST), Graduate School of Culture Technology;Department of Industrial Engineering, Korea Advanced Institute of Science and Technology (KAIST), Daejeon, Korea;Department of Industrial Engineering, Korea Advanced Institute of Science and Technology (KAIST), Daejeon, Korea
Venue:
ICIC'06 Proceedings of the 2006 international conference on Intelligent Computing - Volume Part I
Year:
2006

Citing 8
Cited 0

Toward memory-based reasoning

Communications of the ACM - Special issue on parallelism
A Weighted Nearest Neighbor Algorithm for Learning with Symbolic Features

Machine Learning
Discriminant Adaptive Nearest Neighbor Classification

IEEE Transactions on Pattern Analysis and Machine Intelligence
A Review and Empirical Evaluation of Feature Weighting Methods for aClass of Lazy Learning Algorithms

Artificial Intelligence Review - Special issue on lazy learning
Weighting Features

ICCBR '95 Proceedings of the First International Conference on Case-Based Reasoning Research and Development
MBNR: Case-Based Reasoning with Local Feature Weighting by Neural Network

Applied Intelligence
Hybrid system of case-based reasoning and neural network for symbolic features

DaWaK'05 Proceedings of the 7th international conference on Data Warehousing and Knowledge Discovery
A hybrid approach of neural network and memory-based learning to data mining

IEEE Transactions on Neural Networks

Quantified Score

Hi-index	0.00

Visualization

Abstract

Case-based reasoning (CBR) is frequently applied to data mining with various objectives. Unfortunately, it suffers from the feature weighting problem. In this framework, similar case retrieval plays an important role, and the k-nearest neighbor (k-nn) method or its variants are widely used as the retrieval mechanism. However, the most important assumption of k-nn is that all of the features presented are equally important, which is not true in many practical applications. Many variants of k-nn have been proposed to assign higher weights to the more relevant features for case retrieval. Though many feature-weighted variants of k-nn have been reported to improve its retrieval accuracy on some tasks, few have been used in conjunction with the neural network learning. We propose CANSY, a feature-weighted CBR with neural network for symbolic features.