Bøger af Zhi-Hua Zhou

Ensemble methods that train multiple learners and then combine them to use, with \textit{Boosting} and \textit{Bagging} as representatives, are well-known machine learning approaches. An ensemble is significantly more accurate than a single learner, and ensemble methods have already achieved great success in various real-world tasks.

This book constitutes the refereed proceedings of the 11th Pacific-Asia Conference on Knowledge Discovery and Data Mining, PAKDD 2007, held in Nanjing, China, May 2007. It covers new ideas, original research results and practical development experiences from all KDD-related areas including data mining, machine learning, data warehousing, data visualization, automatic scientific discovery, knowledge acquisition and knowledge-based systems.

The Pacific Rim International Conference on Artificial Intelligence (PRICAI) is one of the preeminent international conferences on artificial intelligence (AI). PRICAI 2008 (http://www.jaist.ac.jp/PRICAI-08/) was the tenth in this series of biennial int- national conferences highlighting the most significant contributions to the field of AI. The conference was held during December 15-19, 2008, in the beautiful city Hanoi, the capital of Vietnam. As in previous years this year's technical program saw very high standards in both the submission and paper review process, resulting in an exciting program that reflects the great variety and depth of modern AI research. This year's contributions covered all traditional areas of AI, including AI foundations, knowledge representation, knowledge acquisition and ontologies, evolutionary computation, etc., as well as va- ous exciting and innovative applications of AI to many different areas. There was particular emphasis in the areas of machine learning and data mining, intelligent agents, language and speech processing, information retrieval and extraction.

This book constitutes the thoroughly refereed revised selected papers from the Second IAPR International Workshop, PSL 2013, held in Nanjing, China, in May 2013. The 10 papers included in this volume were carefully reviewed and selected from 26 submissions. Partially supervised learning is a rapidly evolving area of machine learning. It generalizes many kinds of learning paradigms including supervised and unsupervised learning, semi-supervised learning for classification and regression, transductive learning, semi-supervised clustering, multi-instance learning, weak label learning, policy learning in partially observable environments, etc.

Part 3 (Chapters 11-16) introduces some advanced topics, covering feature selection and sparse learning, computational learning theory, semi-supervised learning, probabilistic graphical models, rule learning, and reinforcement learning.

In closing, Part IV addresses the development of evolutionary learning algorithms with provable theoretical guarantees for several representative tasks, in which evolutionary learning offers excellent performance.

This self-contained introduction shows how ensemble methods are used in real-world tasks. It first presents background and terminology for readers unfamiliar with machine learning and pattern recognition. The book then covers the main algorithms and theories, including Boosting, Bagging, Random Forest, averaging and voting schemes, and diversity measures. Moving on to more advanced topics, the author explains details behind ensemble pruning and clustering ensembles. He also describes developments in semi-supervised learning, active learning, cost-sensitive learning, class-imbalance learning, and comprehensibility enhancement.