En savoir plus sur le livre
This book presents empirical methods for studying complex computer exploratory tools aimed at discovering data patterns, designing experiments, and testing hypotheses to make data more persuasive. Unlike other sciences, computer science and artificial intelligence lack a dedicated curriculum in research methods. The text emphasizes empirical methods, particularly in the context of broader empirical research rather than solely focusing on statistical techniques. The initial chapters introduce empirical questions, exploratory data analysis, and experiment design, while a critical examination of statistical hypothesis testing is addressed in later chapters, which cover classical parametric methods and Monte Carlo resampling techniques. The book is notable for its accessible presentation of these flexible resampling methods. It also emphasizes research strategies and tactics through case studies. Subsequent chapters delve into performance assessment, identifying interactions and dependencies among factors affecting performance, and discussing predictive and causal models. The concluding chapter explores the nature of theory in AI and how empirical methods can contribute to general theories. Mathematical details are provided in appendices, and no prior knowledge of statistics is required. Examples can be analyzed manually or with standard statistics software. The Common Lisp Analytical Statistics Package (CLASP) is available from T
Achat du livre
Empirical Methods for Artificial Intelligence, Paul R. Cohen
- Langue
- Année de publication
- 1995
- Reliure
- (rigide)
Modes de paiement
Personne n'a encore évalué .
- Titre
- Empirical Methods for Artificial Intelligence
- Langue
- Anglais
- Auteurs
- Paul R. Cohen
- Éditeur
- MIT Press
- Publié
- 1995
- Format
- rigide
- Pages
- 405
- ISBN10
- 0262032252
- ISBN13
- 9780262032254
- Séries
- Mots clés
- Nonfiction, Technologie & Ingénierie, Informatique & Internet, États-Unis, Recherche, Intelligence Artificielle, Méthodologie
- Description
- This book presents empirical methods for studying complex computer exploratory tools aimed at discovering data patterns, designing experiments, and testing hypotheses to make data more persuasive. Unlike other sciences, computer science and artificial intelligence lack a dedicated curriculum in research methods. The text emphasizes empirical methods, particularly in the context of broader empirical research rather than solely focusing on statistical techniques. The initial chapters introduce empirical questions, exploratory data analysis, and experiment design, while a critical examination of statistical hypothesis testing is addressed in later chapters, which cover classical parametric methods and Monte Carlo resampling techniques. The book is notable for its accessible presentation of these flexible resampling methods. It also emphasizes research strategies and tactics through case studies. Subsequent chapters delve into performance assessment, identifying interactions and dependencies among factors affecting performance, and discussing predictive and causal models. The concluding chapter explores the nature of theory in AI and how empirical methods can contribute to general theories. Mathematical details are provided in appendices, and no prior knowledge of statistics is required. Examples can be analyzed manually or with standard statistics software. The Common Lisp Analytical Statistics Package (CLASP) is available from T