Machine Learning
Introduction to Algorithms
A Memory-Based Approach to Anti-Spam Filtering for Mailing Lists
Information Retrieval
Text Categorization with Suport Vector Machines: Learning with Many Relevant Features
ECML '98 Proceedings of the 10th European Conference on Machine Learning
The Role of Information Extraction for Textual CBR
ICCBR '01 Proceedings of the 4th International Conference on Case-Based Reasoning: Case-Based Reasoning Research and Development
A study of factors affecting the utility of implicit relevance feedback
Proceedings of the 28th annual international ACM SIGIR conference on Research and development in information retrieval
A survey on tree edit distance and related problems
Theoretical Computer Science
Supervised latent semantic indexing using adaptive sprinkling
IJCAI'07 Proceedings of the 20th international joint conference on Artifical intelligence
A propositional approach to textual case indexing
PKDD'05 Proceedings of the 9th European conference on Principles and Practice of Knowledge Discovery in Databases
Measuring the complexity of a collection of documents
ECIR'06 Proceedings of the 28th European conference on Advances in Information Retrieval
Robust Measures of Complexity in TCBR
ICCBR '09 Proceedings of the 8th International Conference on Case-Based Reasoning: Case-Based Reasoning Research and Development
Introspective knowledge revision in textual case-based reasoning
ICCBR'10 Proceedings of the 18th international conference on Case-Based Reasoning Research and Development
Hi-index | 0.00 |
This paper deals with two relatively less well studied problems in Textual CBR, namely visualizing and evaluating complexity of textual case bases. The first is useful in case base maintenance, the second in making informed choices regarding case base representation and tuning of parameters for the TCBR system, and also for explaining the behaviour of different retrieval/classification techniques over diverse case bases. We present an approach to visualize textual case bases by "stacking" similar cases and features close to each other in an image derived from the case-feature matrix. We propose a complexity measure called GAME that exploits regularities in stacked images to evaluate the alignment between problem and solution components of cases. GAMEclass, a counterpart of GAME in classification domains, shows a strong correspondence with accuracies reported by standard classifiers over classification tasks of varying complexity.