A PREDICTION APPROACH BASED ON ARTIFICIAL NEURAL NETWORKS WITH CONSIDERATION OF ENVIRONMENTAL AND ECONOMIC INDICATORS FOR CAR SALES IN TURKEY

Journal Title: Endüstri Mühendisliği - Year 2015, Vol 26, Issue 3

Abstract

In this study, car sales forecasting for different segments and brands has been aimed considering both environmental and economic indicators. For this purpose, segments and brands of products, past sales quantity, interest rate, gross national product, CO2 emission, fuel consumption and fuel price have been selected as prediction inputs. Using those prediction inputs, artificial neural network model has been recommended to predict car sales. Dataset which is used for prediction, includes car sales of three different brands and three different segments of each brand between the years of 2008-2012. For the artificial neural network model, feed neural network model (FFNN) has been applied. Levenberg-Marquadt algorithm has been used for training of the model. Results of artificial neural networks and linear regression have been compared with each other. Ultimately, It has been acquired that artificial neural network model has more accurate prediction results than linear regression model.

Authors and Affiliations

Yusuf KUVVETLİ, Cansu Dağsuyu, Murat OTURAKÇI

Keywords

Related Articles

A LINE FEEDING SYSTEM DECIDING AGV NUMBERS AND STOCK AMOUNT NEAR LINES

In a company operating in the durable consumer goods sector and producing washing machines, the materials are kept in an interim stock area called supermarket before being sent to the assembly lines. Since there is no sp...

LEAN SIX SIGMA APPROACH AND AN APPLICATION FOR DEFENCE INDUSTRY

Six sigma approach is one of the most important approaches to compete with ever increasing quality expectations in the manufacturing and services sectors. In the context of this study, a lean six sigma literature researc...

State Feedback H∞ Semi Active Control of Fighter Jet Seat

This paper deals with the design of a Linear Matrix Inequalities (LMI) based state feedback H∞ controller for a semi active vibration mitigation problem of fighter jet seats. Ten degrees of freedom detailed full aircraft...

EXACT SOLUTION APPROACHES FOR THE DIRECTED BI-OBJECTIVE CHINESE POSTMAN PROBLEM

In this study, we consider a directed bi-objective Chinese Postman Problem with two additive objectives (like total cost and total distance) and propose two solution approaches to generate all non-dominated objective vec...

An Experimental Investigation of Saturated Flow Boiling Characteristics in Square Microchannels

In this study, saturated flow boiling characteristics of deionized water are experimentally investigated in silicon heat sink consisting of twenty nine parallel microchannels with cross sectional dimensions of 200 µm x 2...

Download PDF file
  • EP ID EP632032
  • DOI -
  • Views 39
  • Downloads 0

How To Cite

Yusuf KUVVETLİ, Cansu Dağsuyu, Murat OTURAKÇI (2015). A PREDICTION APPROACH BASED ON ARTIFICIAL NEURAL NETWORKS WITH CONSIDERATION OF ENVIRONMENTAL AND ECONOMIC INDICATORS FOR CAR SALES IN TURKEY. Endüstri Mühendisliği, 26(3), 23-31. https://europub.co.uk/articles/-A-632032