This paper considers some approaches to building a regression model and a seasonal autoregressive (moving average) integrated model using the Python programming language. The additive regression model was created by using Facebook's Prophet library. The seasonal integrated autoregressive model was created by using the StatsModels library. We developed a prognostic time series of the monthly precipitation sum for the next 2 years. Program experiments were conducted by using data acquired on a Tomsk station (station synoptic index 29430) with an observation period from 1996 to 2016. An interactive environment called Jupiter Notebook was used for the initial data processing, mathematical calculations, and graph plotting. The environment in question is a graphical web-interface for Python which expands the idea of console approach for interactive computing. The model prediction accuracy was assessed by finding the absolute and average absolute errors. The maximum values of the studied time series could not be predicted.
|Journal||IOP Conference Series: Earth and Environmental Science|
|Publication status||Published - 17 Dec 2018|
|Event||International Conference and Early Career Scientists School on Environmental Observations, Modeling and Information Systems, ENVIROMIS 2018 - Tomsk, Russian Federation|
Duration: 5 Jul 2018 → 11 Jul 2018
ASJC Scopus subject areas
- Environmental Science(all)
- Earth and Planetary Sciences(all)