Time Series Forecasting for Non-stationary Data: A Case Study of Petrochemical Product Price

The packaging industry is a dynamic sector of industry that is projected to grow by almost 3% per year for the next ten years. The plastic packaging industry that contributes to the growth of the packaging industry among other packaging materials has a ten-year compound annual growth rate of almost 5%, according to a 2016 report. While plastic is derived from oil, the raw material of plastic is a petrochemical product which is called resin. As resin price fluctuates with the price of global energy, demand and supply, macro-economics, etc, plastic converters as the ones converting resin to plastic products may suffer from the difficulty to pass the raw material price into the product price for customers. Based on the importance of a price forecast for resin as well as limited research in the area of petrochemical product price forecasting (except for the forecasting of oil price itself as the source of the petrochemical products), this paper aims to conduct time series forecasting on the price of resin as one of the petrochemical products. As time series forecasting ranges from traditional method to artificial intelligence method, this paper provides results from using neural network as an artificial intelligence method and the comparison to a traditional method, ARIMA. The result shows that the forecasting accuracy by using ARIMA is higher compared to NN for this particular resin price data. Future research is suggested to seek a time series forecasting method that can best represent the characteristic of resin price in terms of forecast accuracy.

[1]  Il Moon,et al.  Forecasting Naphtha Price Crack Using Multiple Regression Analysis , 2012 .

[2]  R. Geyer,et al.  Production, use, and fate of all plastics ever made , 2017, Science Advances.

[3]  Zhe Lin,et al.  Modelling and forecasting the stock market volatility of SSE Composite Index using GARCH models , 2018, Future Gener. Comput. Syst..

[4]  Hemerson Pistori,et al.  Long-term forecast of energy commodities price using machine learning , 2019, Energy.

[5]  R. Ferdiana,et al.  Application of ARCH model on nutmeg price forecasting in South Aceh district , 2019, IOP Conference Series: Materials Science and Engineering.

[6]  Robert Goodell Brown,et al.  Practical Business Forecasting , 1987 .

[7]  Forecasting ethanol market volatility: new evidence from the corn implied volatility index , 2018, Biofuels, Bioproducts and Biorefining.

[8]  M. Zwicky,et al.  Mapping of global plastic value chain and plastic losses to the environment : with a particular focus on marine environment , 2018 .

[9]  David S. Stoffer,et al.  Time series analysis and its applications , 2000 .

[10]  Robert H. Shumway,et al.  Time series analysis and its applications : with R examples , 2017 .

[11]  Mansur Masih,et al.  Price dynamics of crude oil and the regional ethylene markets , 2010 .

[12]  M. Hashem Pesaran,et al.  A Simple Nonparametric Test of Predictive Performance , 1992 .

[13]  Aderemi Oluyinka Adewumi,et al.  Comparison of ARIMA and Artificial Neural Networks Models for Stock Price Prediction , 2014, J. Appl. Math..

[14]  Murat Kulahci,et al.  Introduction to Time Series Analysis and Forecasting , 2008 .

[15]  Gan-qiong Li,et al.  Short-Term Price Forecasting For Agro-products Using Artificial Neural Networks , 2010 .

[16]  Gautam Bandyopadhyay,et al.  Gold Price Forecasting Using ARIMA Model , 2015 .

[17]  F. Diebold,et al.  Comparing Predictive Accuracy , 1994, Business Cycles.

[18]  H. Mohammadi,et al.  International evidence on crude oil price dynamics: Applications of ARIMA-GARCH models , 2010 .

[19]  Ani Shabri,et al.  Daily crude oil price forecasting model using arima, generalized autoregressive conditional heteroscedastic and Support Vector Machines , 2014 .