摘要：database, the use of Markov chain models is presented to predict a ‘Snow Day’ at eight national weather stations. This serves as a variation of the classic Markov chain precipitation example, predicting a significant snow depth tomorrow from today’s snow depth conditions. Stations near Seattle WA, Denver CO, Milwaukee WI, Chicago IL, New York NY and Boston MA, were included as they represent major urban centers, while stations in Montana and North Dakota were added to improve geographical coverage. Estimates of the appropriate transition matrices (P.i) are provided, as well as a sample of code in the R statistical programming language to enable construction of similar examples for other geographical areas.
