A hybrid SOM-Fuzzy time series (SOMFTS) technique for future forecasting of COVID-19 cases and MCDM based evaluation of COVID-19 forecasting models

Kumar, A.; Kaur, K.

This paper proposes a hybrid technique based on self-organized maps and fuzzy time series (SOMFTS) for future forecasting of COVID-19 cases. This paper also presents an approach for evaluation of COVID-19 forecasting models based on Multi Criteria Decision Making (MCDM). Since the evaluation of forecasting models involves more than one performance measures, it can be modeled as an MCDM problem. The experimental study presented in this paper evaluates the proposed new SOMFTS technique and seven conventional COVID-19 forecasting techniques. The results of this paper demonstrate the efficiency of SOMFTS technique for future forecasting of COVID -19 cases and the utility of MCDM methods for evaluation and selection of COVID-19 forecasting models. To demonstrate our proposed SOMFTS forecasting technique and MCDM based approach for evaluation and selection COVID-19 forecasting models, we take the number of confirmed, cured and death cases in Delhi, India, as a case study.