Information and scientific discovery: Shannon, Bayes and Leydesdorff
摘要
Loet Leydesdorff pioneered in the application of Shannon’s information theory and Bayes’ theorem to scientometric data. He focused on word distributions in full texts and developing methods for both cross-sectional analysis and change over time in information theory terms. As an extension, or perhaps a deviation from his approach, I present an information theory treatment of theory-evidence agreement, showing how Shannon’s ideas on entropy can be applied to understanding scientific discovery and its relation to Bayesian confirmation.