PREDICTION FOR SECOND HAND CARS SELLING ON THE INTERNET WITH THE ARTIFICIAL NEURAL NETWORK METHOD

Abstract

Data mining is an important field of study for e-commerce firms. With the overwhelming amount of available data, that increases exponentially it is inevitable to use data mining methods. Because cost of a good or a service, potential of selling the product, potential price of the product etc. can be estimated by these methods with enough variables. No matter which method we use it must be cost and time efficient. For this reason, we must use data mining methods to make accurate and immediate estimations without any delays. In this work we tried to estimate prices of the vehicles in the secondhand car market with artificial neural network method. For this we used six step process to solve this problem. Problem definition was made by using data mining stages, in the first stage data cleaning was done in data preparation, in second stage we arranged data for the exploration, in third stage modelling was done, in fourth stage created model was evaluated, and in the fifth step data was adapted by model deployment to the working principles of the algorithms that would be used. Then, in the final stage it was evaluated by the methods of multilayer perceptron aka artificial neural network method. Results from artificial neural networks method compared with the actual data and the results analyzed.