Prediction of O3 Concentration Level Using Fuzzy Non-Stationary Method

Abstract

The composition of air concentration is not constant. It constantly changes with minor changes at any time, so more than one measurement is needed to represent the air concentration level for a full day. The fuzzy non-stationary method can overcome uncertainty in an environment that is not constant or caused by minor temporal changes based on time variables. This study uses a non-stationary fuzzy method to determine the level of O3 concentration based on the input variables of temperature, humidity, and wind speed. The tests were conducted in September, October, and November using four types of implication process interpretation, namely interpretation 1 (classical logic), interpretation 2 (classical logic), interpretation 3 (algebraic), and interpretation 3 (standard). The test results in September showed a tendency for error percentage using the MAPE amount of 19, October's amount of 25, and November's amount of 18.