The Philippines, like other developing countries, relies on borrowings to increase its meager resources. This study aims to predict the national outstanding debt of the Philippines from December 2018 to December 2023 and compare the two methods, specifically, multiple linear regression and feed forward backpropagation neural network for which is the best model to forecast the national outstanding debt. The data used were gathered through Philippine Statistics Authority and BangkoSentralngPilipinas. Multiple linear showed all the eight independent variables namely inflation rate , government revenue , government expenditure , exchange rate , import , export , government surplus , and domestic interest rate have a significant relationship with national outstanding debt , but only government revenue , exchange rate , import , government surplus , domestic interest rate have a significant impact to the national debt. Any movement on these said variables can cause the national outstanding debt to be inclined or declined. However, measuring accuracy formulas were also applied in multiple linear regression and feed forward backpropagation neural network. And the results obtained were considered to compare which model may give the prediction of national outstanding debt more accurately. The results revealed that feed forward backpropagation neural network has a least percentage error than multiple linear regression. Thus, a best fitted model on predicting national outstanding debt. This study can show an overview on the past, current, and future condition of Philippines’ debt.
Volume 12 | 06-Special Issue
Pages: 336-353
DOI: 10.5373/JARDCS/V12SP6/SP20201040