LTKlasikiniai ekonometrijos modeliai tampa neefektyvūs kuomet tiriami erdviniai duomenys. Tokiais atvejais siūloma rinktis alternatyvius modelius, atsižvelgiančius į erdvines duomenų charakteristikas. Straipsnyje apžvelgiami du erdvinės regresijos modeliai bei pateikiamas šių modelių taikymo pavyzdys modeliuojant Lietuvos bendrąjį vidaus produktą (BVP). Sudarytas modelis lyginamas su klasikinės regresijos modeliu. Gauti rezultatai leidžia daryti išvadą, kad modelis, sudarytas įtraukiant erdvinę informaciją, yra statistiškai reikšmingai geresnis. [Iš leidinio]Reikšminiai žodžiai: Bendrasis vidaus produktas; Erdviniai duomenys; Erdvinio vėlavimo modelis; Erdvinių klaidų modelis; Makroekonominiai rodikliai; Gross domestic product; Macroeconomic indicators; Spatial data; Spatial lag model; Spatial regression model.
ENClassical econometrics becomes ineffective when spatial data are being studied. In such cases, it is proposed to use alternative methods that take into account spatial information. The paper discusses two spatial regression models: spatial lag model and spatial error model. In order to demonstrate the advantage of these models against the classical regression, the example of modeling of Lithuanian gross domestic product (GDP) is provided. Four macroeconomic indicators as explanatory variables are used: an unemployment rate, average monthly income, foreign direct investment and the number of working population. The data were collected from Lithuanian counties in 2017. Since the hypothesis about spatial correlation was accepted a spatial lag model was built to model GDP. The proposed model is compared to the classic regression model. The results obtained suggest that the model with included spatial information is statistically significantly better than the model without this component. [From the publication]