Probabilistic Causal Knowledge Learning for Everyday Life Behavior Computing

摘要

Bayesian networks and statistical learning methods are studied for everyday life computational processing with real world sensor data. In particular, graphical models that have complex structures can handle prior knowledge involved in target problems. Recently, application of graphical models like Bayesian networks are reported as efficient modeling for real world problems such as pattern recognition and time series data analysis. In this paper, we apply Bayesian network learning to model causal structure in everyday life behavior and show Bayes estimation of image understanding with naive Bayes classifiers.