A novel discrete grey seasonal model and its applications

In order to accurately describe real systems with seasonal disturbances, which normally appear monthly or quarterly cycles, a novel discrete grey model, abbreviated as DGSM(1, 1), is put forward by incorporating the dummy variables into conventional model. Moreover, mechanism and properties of this proposed model are discussed in depth, revealing inherent differences from existing models. For validation explanation purposes, implemented three actual cases fluctuations (quarterly wind power production, PM10, natural gas consumption), comparison five competing models involving prediction models, econometric technology, artificial intelligences. Experimental results consistently demonstrate that significantly outperforms other benchmark terms several error criteria. further discussions about influences different sequence lengths on forecasting performance reveal still performs best strong robustness high reliability addressing sequences. general, new validated be powerful promising methodology for handling sequences fluctuations.