Recent Trends in Interval Regression: Applications in Predicting Dengue Outbreaks

Dengue disease is a serious threat for the world. The number of infections increase annually forcing implementation of prompt actions in dengue management. Common practice of modelling is associated with point measurements. However, an interval representation for a point measure provides an additional information for the spread, capture uncertainties associated with variables and useful in making more precise decisions. Further, interval predictions are appropriate in the situations of exact predictions are not essential. Interval-valued analysis in the dengue disease is important as actions taking towards controlling the disease do not depend on the exact number but on the magnitude of the values represented by the interval. In the area of regression analysis, there are techniques to handle interval-valued dependent and independent variables. The present chapter discusses theories of interval regression procedures: centre method, centre and range method, constrained centre and range method, interval regression based on interval least squares algorithm and fuzzy regression techniques. The chapter illustrates applications of these methods using interval-valued data in Colombo, Sri Lanka, and Jakarta, Indonesia. Finally, the chapter emphasizes the importance and effectiveness of the interval regressions over traditional linear regression as well as added advantages of soft computing methods.