Open AccessApplication of Neural Network Technologies for Inflation Forecasting
Publication year - 2020
Publication title -
international journal of mathematical models and methods in applied sciences
Language(s) - English
Resource type - Journals
ISSN - 1998-0140
DOI - 10.46300/9101.2020.14.9
Subject(s) - inflation (cosmology) , artificial neural network , smoothing , econometrics , computer science , autoregressive integrated moving average , receipt , exponential smoothing , time series , technology forecasting , quality (philosophy) , economics , artificial intelligence , machine learning , philosophy , physics , epistemology , theoretical physics , world wide web , computer vision
The paper considers the methodology of forecasting the level of inflation in Russia with the help of analytical platform Deductor Studio. In solving the problem, the mechanisms of data purification from noises and anomalies were applied, which ensured the quality of forecast model construction and receipt of forecast values for five months in advance. The principle of forecasting the time series was also demonstrated: import, seasonal detection, cleaning, smoothing, construction of forecast model, and forecasting the inflation rate for five months ahead.