Open AccessPredicting the Albanian economic development using multivariate Markov chain model
Author(s) -
Eralda Gjika,
Lule Basha,
Xhensilda Allka,
Aurora Ferrja
Publication year - 2020
Language(s) - English
Resource type - Conference proceedings
DOI - 10.3846/bm.2020.581
Subject(s) - multivariate statistics , markov chain , life expectancy , inflation (cosmology) , econometrics , economic indicator , unemployment , markov model , index (typography) , computer science , time series , multivariate analysis , economic forecasting , statistics , economics , mathematics , demography , economic growth , macroeconomics , population , physics , sociology , theoretical physics , world wide web
In this work, the economic development and relation to social and demography indices in Albania were studied. Four time series (yearly data for the period 1995–2018) were considered: consumer price index (CPI), unemployment rate, inflation and life expectancy. In our approach, a first and fifth order multivariate Markov chain model was proposed to predict the economic situation in Albania in the proceedings years. Tests and accuracy analysis of the model were performed. The prediction probabilities fall in the interval of 0.47 to 0.52 and the accuracy of both models is 75%. Our approach is a short term probability forecast model that can be used by the policymakers to evaluate and undertake initiatives to improve the situation in the country.