As described in , time series data includes many kinds of real experimental data taken from various domains such as finance, medicine, scientific research (e.g., global warming, speech analysis, earthquakes), etc. Time series forecasting has many real applications in various areas such as forecasting of business (e.g., sales, stock), weather, decease, and others . Statistical modeling and inference (e.g., ARIMA model)  is one of the popular methods for time series analysis and forecasting. The philosophy of Bayesian inference is to consider probability as a measure of believability in an event  and use Bayes' theorem to update the probability as more evidence or information becomes available, while the philosophy of frequentist inference considers probability as the long-run frequency of events . Generally speaking, we can use the Frequentist inference only when a large number of data samples are available.
Jul-5-2020, 18:33:26 GMT