| Abstract [eng] |
The aim of this master‘s project is to estimate time delay parameter between financial time series and to find the best forecasting model with included delay parameter. The research object – Lithuanian general leasing portfolio. In the first part of project, the time delay estimation was computed in comparison with financial time series, such as inflation, consumer price index, export, import, the number of immigrants and the unemployment rate. The results has showed that estimated time delay is very important for the evaluation of economics and its relations. In the second part of project,the results of Granger causality test has showed that Lithuanian general leasing portfolio is influenced by inflation. Time delay parameter was analysed between those time series with cross correlation, second correlation and multilayer correlation methods. Forecasting of leasing portfolio is performed using Autoregressive Distributed Lag (ADL) model (which compared time delay parameter), Autoregressive Integrated Moving Average (ARIMA) model and Vector Autoregression (VAR) model. The mean absolute percentage error (MAPE) was calculated to evaluate the accuracy of the future forecasts and then the validation was carried out between the actual and predicted values. Models are implemented in Matlab and SAS software. |
