Abstract:
Crude oil plays a crucial role as a commodity with
significant global economic implications. This research project
aimed to identify predictive models for forecasting future oil
prices in Sri Lanka, utilizing monthly data spanning over 32
years. The analysis involved examining monthly Domestic Prices
for five petroleum products, Lanka Petrol 95 (LP95), Lanka
Petrol 92 (LP92), Lanka Auto Diesel (LAD), Lanka Super Diesel
(LSD), and Lanka Kerosene (LK) covering the period from
1990 to 2021. Three popular time series trend models, namely,
Linear Trend Model (LTM), Quadratic Trend Model (QTM)
and Exponential Growth Model (EGM), were used on the five
petroleum product types. Furthermore, the fitted models were
further assessed to find the best fitted models for each type
of petroleum product using accuracy measures such as Mean
Squared Deviation (MSD), Mean Absolute Deviation (MAD),
and Mean Absolute Percentage Error (MAPE). According to the
trend analysis results, the exponential growth model was the
most suitable for LP95, LP92, LAD, LSD and LK. The results
of the study offer practical implications for stakeholders in the
Sri Lankan petroleum industry, enabling them to make informed
decisions in a volatile global market.
