A Testbed for Experiments in Adaptive Memory Retrieval and Indexing

Liwu Chang & Patrick Harrison

In this paper, we discuss properties of the Knack system which is designed as a testbed for experimenting memory retrieval and organization in Case-Based Reasoning. Methods used for retrieval and indexing are based on mathematically sound techniques developed in classification, clustering and decision analysis. Retrieval is done using decision theoretic methods such as voting k nearest neighbor and Bayesian theory with weighted attributes. New indices for cases are generated by using clustering methods. Cases are then re-organized using the new indices. The Knack environment was designed so that additional techniques and metrics can easily be added.

This page is copyrighted by AAAI. All rights reserved. Your use of this site constitutes acceptance of all of AAAI's terms and conditions and privacy policy.