Analysis of The Application of Fuzzy Logic and Levenberg-Marquardt in The Calculation of Used Car Prices

Abstract

In this study using two methods, namely Sugeno Fuzzy Logic and Levenberg-Marquartd methods. The data to be used is data from used Toyota Avanza cars for Medan and surrounding areas. The use of the fuzzy method in this study is based on its superiority in solving vague problems so that it is considered to be suitable for the condition of a used car in which each unit has different conditions. The results of the use of Sugeno fuzzy data obtained will be used as input into the levenberg-marquardt (LM) network which has a fairly good ability to perform calculations, especially to predict an output with a more efficient time. LM produces an output with a small Minimun Square Error (MSE) value of 3.64 in testing with 9 hidden layers. On the other hand, to analyze the resulting error, a comparison is made between networks that have 9 hidden layers, 27 hidden layers and 81 hidden layers with the smallest MSE results obtained on a network with 9 hidden layers.