An Interval-Valued Time Series Forecasting Scheme With Probability Distribution Features for Electric Power Generation Prediction

Abstract

Developing an effective interval-valued time series (ITS) forecasting scheme for electric power generation is an important issue for energy operators and governments when making energy strategic decisions. The existing studies for ITS forecasting only consider basic descriptive information such as center, radius, upper and lower bounds, and overlooks the distribution information within the data interval. In this study, an interval-valued time series forecasting scheme based on probability distribution information features of interval-valued data with machine learning algorithms is proposed to enhance electric power generation forecasting. In the proposed scheme, the central tendency features and dispersion features from the interval-valued data are designed as integrated features sets (IFS) and used as predictor variables. Three methods including supper vector regression and extreme learning machine and multivariate adaptive regression splines based on the IFS are utilized to develop ITS forecasting models. The daily time series of the metered generation from the Australian Energy Market Operator is used to illustrate the proposed scheme. Empirical results show that the proposed ITS forecasting schemes with IFS outperform the eight benchmark models and thus validate that the proposed scheme is an effective alternative for interval-valued electric power generation forecasting.

Interval time series forecasting: A systematic literature review

“…To efficiently use the potential information contained in interval data, Han and Wang (2012), Han et al (2016), Wang et al (2016), Sun, Han, et al (2018), Sun, Zhang, et al (2019), Sun et al (2021), Sun, Bao, et al (2020) and He et al (2021) established a set‐based interval regression method. Wu and Perloff (2005), Arroyo et al (2010), González‐Rivera and Lin, (2013), Lin and González‐Rivera (2016), Gonzalez‐Rivera et al (2020), and Chang et al (2022) have proposed inference methods for interval‐valued regression to analyze the interval distribution of random variables. González‐Rivera and Arroyo (2012); Golan and Ullah (2017); Buansing et al (2020); Guo et al (2021); and Hao et al (2022) developed information‐theoretical methods for interval analysis.…”

“…Information‐source‐based studies are aimed at improving the predictability of the interval time series by supplementing external information sources (Carvalho & Martos, 2021; Hajek et al, 2020; Xiong et al, 2018, 2020). These studies primarily include the use of historical models, influencing‐factor‐based models (Chang et al, 2022; Garda‐Martos et al, 2013; Huang, Sun, & Wang, 2021; Ji et al, 2022; Lu et al, 2022; Tan & Wang, 2017; Wang, Gao, et al, 2023; Xu & Qin, 2023), and unstructured models (Amuri & Marcucci, 2017; Fan et al, 2017; Huang & He, 2020; Ren et al, 2018; Sun et al, 2017; Sun, Wei, et al, 2019). These studies suggest that dynamic forecasting systems can be developed using forecasting models based on the influencing variables and unstructured data.…”

An Experimental Machine Learning Approach for Mid-Term Energy Demand Forecasting

“…The scheduling of repairs and maintenance, as well as the efficient use of the power plants, which are closely tied to the reliability, are the primary purposes of MTLF, which has gained significant prominence as a crucial instrument in the last ten years. The MTLF horizon is from several months to 1,2 years [10]. The MTLF assists with resource allocation throughout this time frame and developing various sub-structural components that may be feasible in the midterm horizon.…”

Interval-valued Data Analysis: A Review

“…However, new kinds of data are arisen called "Interval data". This data is a typical interval with lower and upper bounds such as weekly highs and lows in daily temperatures, heart tension values and many other magnitudes that do not give the variables as a single value as we are used to [7][8][9].…”