SutteARIMA is a New Approach to Forecast Economics, Business, and Actuarial Data

Abstract

[eng] The main objective of this study was to develop a new forecasting method i.e. SutteARIMA method. SutteARIMA was developed by using a combination and/or weaknesses of some forecasting methods that already exist (α-Sutte Indicator and ARIMA). Based on the this main objective, it will be outlined into three specific objectives as follows: (1) development of new forecasting method (SutteARIMA) in the sector of finance; (2) forecasting of financial and actuarial data by using SutteARIMA method; (3) comparative study of financial data forecasting results. The specific objectives of this research have been discussed and implemented so as to produce several general conclusions, namely: SutteARIMA method has a better level of accuracy for predicting economics, business, and actuarial data which has been discussed in Chapters 2, 3, and 4. The results of this conclusion are strengthened by obtaining the mean absolute percentage error (MAPE) and mean squared error (MSE) which is smaller when compared to other forecasting methods.

In Chapter 2, this study focuses on predict the short-term of confirmed cases of covid-19 and IBEX in Spain by using SutteARIMA method. Covid-19 Spanish confirmed data obtained from Worldometer and Spain Stock Market data (IBEX 35) data obtained from Yahoo Finance. Data starts from 12 February 2020 – 09 April 2020 (the date on Covid-19 was detected in Spain). The data from 12 February 2020 – 02 April 2020 using to fitting with data from 03 April – 09 April 2020. Based on the fitting data, we can doing short forecast for 3 future period (10 April – 12 April 2020 for Covid-19 and 14 April – 16 April 2020 for IBEX). In this study, the SutteARIMA method will be used. For the evaluation of the forecasting methods we applied forecasting accuracy measures, mean absolute percentage error (MAPE). Based on the results of ARIMA and SutteARIMA forecasting methods, we conclude that the SutteARIMA method is most suitable than ARIMA to calculate the daily forecasts of confirmed cases of Covid-19 and IBEX in Spain. The MAPE value of 0.1905 (smaller than 0.04 compared to MAPE value of ARIMA) for confirmed cases of Covid-19 in Spain and 0,0202 for IBEX stock. At the end of the analysis, using the SutteARIMA method, we calculate daily forecasts of confirmed cases of Covid-19 in Spain from 10 April 2020 until 12 April 2020 and Spain Stock Market from 14 April until 16 April 2020.

In Chapter 3, this study focuses on the novel forecasting method (SutteARIMA) and its application in predicting Infant Mortality Rate data in Indonesia. It undertakes a comparison of the most popular and widely used four forecasting methods: ARIMA, Neural Networks Time Series (NNAR), Holt-Winters, and SutteARIMA. The data used were obtained from the website of the World Bank. The data consisted of the annual infant mortality rate (per 1000 live births) from 1991 to 2019. To determine a suitable and best method for predicting Infant Mortality rate, the forecasting results of these four methods were compared based on the mean absolute percentage error (MAPE) and mean squared error (MSE). The results of the study showed that the accuracy level of SutteARIMA method (MAPE: 0.83% and MSE: 0.046) in predicting Infant Mortality rate in Indonesia was smaller than the other three forecasting methods, specifically the ARIMA (0.2.2) with a MAPE of 1.21 % and a MSE of 0.146; the NNAR with a MAPE of 7.95% and a MSE of 3.90; and the Holt-Winters with a MAPE of 1.03% and a MSE: of 0.083.

In Chapter 4, this chapter proposes a new model, namely the SutteARIMA model, combining the α-Sutte Indicator and ARIMA methods to forecast economic and finance data regardless of whether the data is linear or non-linear. The proposed model has the following advantages: it does not pay attention to linear-nonlinear data, the forecasting accuracy is more stable, and the calculation rate is faster. To evaluate the performance, SutteARIMA is compared with conventional models provided (linear and non-linear) namely ARIMA, Holt-Winters, Neural Network, Robust, and Theta models. The results of the study showed that the SutteARIMA method is more accurate than other models, based on MSE: 66474.88 / MAPE: 1.33% on National Currency to US Dollar Spot Exchange Rate for Indonesia and MSE: 0.0493 / MAPE: 0.1594% on Consumer Price data Index: All Items for Indonesia. In addition, from the results of research and test results on the generated artificial data, it is also found that SutteARIMA is suitable for Non Trend, Trend, Non Seasonal, Seasonal data, and their combination.