Instance-Based Learning Algorithms
Machine Learning
Trading MIPS and memory for knowledge engineering
Communications of the ACM
An Average-Case Analysis of k-Nearest Neighbor Classifier
ICCBR '95 Proceedings of the First International Conference on Case-Based Reasoning Research and Development
A Symmetric Nearest Neighbor Learning Rule
EWCBR '00 Proceedings of the 5th European Workshop on Advances in Case-Based Reasoning
Average-Case Analysis of Classification Algorithms for Boolean Functions and Decision Trees
ALT '00 Proceedings of the 11th International Conference on Algorithmic Learning Theory
Effects of domain characteristics on instance-based learning algorithms
Theoretical Computer Science - Selected papers in honour of Setsuo Arikawa
| Hi-index | 0.00 |
This paper presents an average-case analysis of the fc-nearest neighbor classifier (k-NN). Our analysis deals with m-of-n// concepts, and handles three types of noise: relevant attribute noise, irrelevant attribute noise, and class noise. We formally compute the expected classification accuracy of fc-NN after a certain fixed number of training instances. This accuracy is represented as a function of the domain characteristics. Then, the predicted behavior of fc-NN for each type of noise is explored by using the accuracy function. We examine the classification accuracy of fc-NN at various noise levels, and show how noise affects the accuracy of fc-NN. We also show the relationship between the optimal value of k and the number of training instances in noisy domains. Our analysis is supported with Monte Carlo simulations.