Data analysis via supervised learning tasks is among the most common data mining techniques. The objective of meta-learning is to generate a user-supporting system for selection of the most appropriate supervised learning algorithms for such tasks. The meta-learning framework is usually based upon a classification on the meta-level often disregarding a large amount of information gained during the induction process. The performance of supervised learning algorithms is also clearly dependent on the quality of the data. And, considering only a small subset of meta-attributes may significantly reduce both the time and effort applied for the corresponding measurement process. In this book, the extent to which the issues above impact the performance of a meta-learning system is evaluated and solutions for remedying the difficulties observed are presented. In particular, the accuracies of the base learners are predicted, thus avoiding the rigid decision on a single-best learner. Subsequently, the severity of data quality issues is investigated.
IOS Press is an international science, technical and medical publisher of high-quality books for academics, scientists, and professionals in all fields.
Some of the areas we publish in:
• Biomedicine
• Oncology
• Artificial intelligence
• Databases and information systems
• Maritime engineering
• Nanotechnology
• Geoengineering
• All aspects of physics
• E-governance
• E-commerce
• The knowledge economy
• Urban studies
• Arms control
• Understanding and responding to terrorism
• Medical informatics
• Computer Sciences