Interval Data Analysis by Revised Interval Regression Model
Hisao Ishibuchi, Hideo Tanaka, Kazunori NAGASAKA · Transactions of the Society of Instrument and Control Engineers · 1989
The main purpose of interval regression is to obtain two interval regression models by which the given interval data are explained in two ways. Formerly the two models were obtained by solving two LP problems called Min and Max problems respectively. Since the two LP problems were independent of each other, the consistency of the two models was not assured. In this paper, we intend to propose the revised method to obtain the two interval regression models being consistent with each other. Our intention is achieved by combining Min problem with Max problem. For the application of the proposed method, two interval regression models representing the relation of the feed speed and the surface roughness in grinding experiments are calculated. Lastly the concept of the interval regression is explained in the context of rough sets and fuzzy sets.