The Artificial Neural Network Application for Service-Oriented Evaluation of the Used Cars

摘要

The article is devoted to the development of a system of the service-oriented valuation to the used cars' cost. "Rapid" produced by the famous Czech manufacturer "Skoda" was chosen as the target brand and model of the car. The paper presents a mathematical formulation of the determining problem of the value of the car in the form of a classification problem. To solve the problem, the authors propose to use the technology of the artificial neural networks (ANN). The universal classifier of the "multilayer perceptron" type with two hidden layers was used as the ANN to be trained. The "NeuroSolutions" neuropackage program is used as the neural network modeling tool. The input and output parameters of the trained ANN are determined. The statistics collection for training the ANN was collected on Russian websites specializing in the sale and purchase of the used cars, in particular, on "auto.ru" and "drom.ru". As a result, the created ANN is able to solve the presented problem with a sufficient degree of accuracy. The article discusses the prospects for the research results' implementation.