Fixed effects panel interval-valued data models and applications

Interval-valued data is a complex data type which can be got by summarizing large datasets, linear regression models for interval-valued data have been widely studied. Panel data models combining cross-section and time series real-valued data have become increasingly popular in economic research and data mining. It is very important to construct the regression models for panel data with uncertainty and range variability. This paper introduces panel data regression model for interval-valued data and constructs three kinds of panel interval-valued data regression models: the centre model of fixed effects panel interval-valued data regression, the min–max model of fixed effects panel interval-valued data regression and its special model, the centre and range model of fixed effects panel interval-valued data regression. Then combining the parameters estimation of interval-valued regression and analysis of covariance for panel data, this paper presents the parameters estimations for three kinds of panel interval-valued data regression models. Finally, our proposed panel interval-valued data regression models are applied in forecasting of Air Quality Index, the experimental evaluation of actual data sets shows the advantages and the performance of our proposed panel interval-valued data models. •Extend fixed effects panel data model from real-valued to interval-valued.•Three models are proposed, including P-CM model, P-Min-Max model, and P-CRM model.•Actual experiments show that the three models have good perform.