Home // DATA ANALYTICS 2016, The Fifth International Conference on Data Analytics // View article
Authors:
Danial Shahrabi Farahani
Mansour Momeni
Nader Sayyed Amiri
Keywords: car sale prediction; analytical hierarchy process; artificial neural networks; feed forward network; multi-layer back propagation neural network; learning algorithm
Abstract:
In this study, we evaluate different effective factors related to marketing and sales and discuss the various prediction methods. The field of this study is the car industry and the tools used for classification, comparison and weight determination is the Analytical Hierarchy Process (AHP). Artificial Neural Networks are used for identifying the architecture and shaping the process of prediction. In order to do so, using a questionnaire presented to experts in the field, the factors affecting car sales in North America were identified and the processed weights obtained from these opinions were fed to the neural network as input, so that, ultimately, by teaching the network through different algorithms, the optimal solution can be obtained. The conceptual model of the research first identifies the factors affecting sales and then tries to determine the interconnection among the data. In order to compare the performance of this method, we needed a valid and established measure so that we can assess the methods based on it. Therefore, linear and exponential regression methods were selected to compare the degree of error and to obtain a more desirable final output which is closer to reality. The obtained result indicates the successful performance of the neural network compared to other selected methods and it was found that it has a lower Minimum Square Error (MSE) compared to others.
Pages: 57 to 62
Copyright: Copyright (c) IARIA, 2016
Publication date: October 9, 2016
Published in: conference
ISSN: 2308-4464
ISBN: 978-1-61208-510-4
Location: Venice, Italy
Dates: from October 9, 2016 to October 13, 2016