A role of causal inference in non-stationary environments

Abstract

Bayesian inference is a process of narrowing down hypotheses (causes) to one that best explains observational data (effects). To accurately estimate a cause, a considerable amount of data is required to be observed for as long as possible. However, the object of inference is not always constant. In this case, a method, such as exponential moving average (EMA) with a discounting rate, is used to improve the ability to respond to a sudden change. That is, a trade-off is established in which the followability is improved by increasing the discounting rate, but the accuracy is lessened. We propose an extended Bayesian inference (EBI), incorporating human-like causal inference into Bayesian inference and evaluate the estimation performance of EBI through the simple numerical experiments on coin toss. The EBI was shown to modify the trade-off seen in EMA.