Intelligent Time Series Forecasting Through Neighbourhood Search Heuristics
Automated forecasting is essential to business operations that handle scores of univariate time series. Practitioners have to deal with thousands of time series with a periodicity ranging from seconds to monthly. The sheer velocity and volume of time seri
- PDF / 839,214 Bytes
- 11 Pages / 439.37 x 666.142 pts Page_size
- 100 Downloads / 270 Views
2
Heriot-Watt University, Edinburgh, UK [email protected] Singapore University of Social Science, Singapore, Singapore [email protected]
Abstract. Automated forecasting is essential to business operations that handle scores of univariate time series. Practitioners have to deal with thousands of time series with a periodicity ranging from seconds to monthly. The sheer velocity and volume of time series make it challenging for human labour to manually identify the order of the time series to forecast the results. An automated forecasting algorithm or framework is essential to complete the task. The approach must be robust in the identification of the order of the time series, and readily applicable to scores of time series without manual intervention. The most modern automated forecasting algorithms are derived from exponential smoothing or ARIMA models. In this paper, the authors proposed a new heuristics approach to identify the initial starting point for a neighbourhood search to obtain the most appropriate model. The results of this method are used to compare against the methods proposed in the literature. Keywords: Data mining ARIMA
Time series Forecasting Heuristics
1 Introduction Automated forecasting for huge volumes of univariate time series is found in many industries. At a single time, there can be easily thousands of product lines that require forecasting on a monthly basis or even shorter time frame. Even when the number of projections required is in the dozens, they still post significant difficulties if there is nobody suitably trained to build the forecasting models. Under such conditions, an automated forecasting approach will be indispensable to the organisation. Automated forecasting strategies must be able to determine the correct or suitable model, estimate the parameters and compute the forecasts. The approach must be robust to handle outliers and does not require manual intervention for most situations. In the healthcare domain, the analysis of time series from clinical sources, operation sources, as well as population information, are keys to management of healthcare demands. Getting an appropriate mechanism or model to forecast demand accurately enables healthcare organisation to refine optimal allocation strategies to satisfy the needs of the patients while minimising inventory costs.
© Springer Nature Switzerland AG 2019 K. Arai et al. (Eds.): FICC 2018, AISC 887, pp. 434–444, 2019. https://doi.org/10.1007/978-3-030-03405-4_30
Intelligent Time Series Forecasting Through Neighbourhood Search Heuristics
435
However, forecasting demand and other time series is a complex problem that requires highly specialised skills and knowledge to complete the task successfully. Retail forecasting requires knowledge of the retail market and consumption of goods. An automatic way to perform forecasting for retail reduces operational costs and time. Accurate forecasts also create opportunities to evaluate the effects of any particular activities or campaigns that can be achieved through comparison
Data Loading...
