Probabilistic inference of environmental factors via time series analysis using mean-field theory of ising model

We constructed a probabilistic modeling forecast variation of temperature using the time-series analysis based on statistical mechanics. In order to clarify availability of the present method, we estimated time evolution of squared error between predictive data and actual data on environmental factors, such as temperature. Using numerical simulation for climate statistics on temperature, we found that the present method was practically useful method via the mean-field theory of the Ising model to forecast environmental factors, although the accuracy of the present method was just inferior to that of the conventional method.