Hybrid Model of Customer Response Modeling Through Combination of Neural Networks and Data Pre-processing Abbas Namdar Aliabadi/MBA Department of Economical and Administrative Science University of Mazandaran Babolsar, Iran a.namdar@umz.ac.ir Hamid Berenji/IEEE Fellow Intelligence Inference Systems Corp MS: 566/108, NASA Research Park Moffett Field, California, 94035 berenji@iiscorp.com Abstract—Due to the increasing data volume, it is difficult for direct marketing decision makers to find target customers. Computational intelligence models that computerize human analysis have recognized a testing device for customer response modeling and target customers identification. Widespread research has resulted in numerous direct marketing applications using computational intelligence models; customer response modeling is an important activity of direct marketing. This paper proposes a hybrid model for customer response modeling. The Proposed model evolves neural network primary weights by genetic algorithms and optimizes it by back-propagation. We also utilize data pre-processing methods such as data reduction for improving precision of the model. We test ability of the proposed method by applying it to standard data set. Results show that the proposed approach is capable to deal with the customer response modeling and it also yields superior prediction accuracy. Keywords—Direct Marketing; Genetic Algorithms; Neural Networks; Data Pre-processing. I. INTRODUCTION The number of people using the internet has enlarged significantly because of the commonly accepted web environment. The internet has also quickly accumulated a huge mass of data and has grown to be one of the greatest means of information storage. In such a web environment, the concept of the direct marketing is interesting as it includes information technology which could generate plenty of complex data. The appearance of direct marketing system storing digitize data makes it possible to search more simply and expediently. Traditionally, the direct marketing system used to play an inactive role in that it only provided customers. It is a critical subject, however, for a marketing manager to think about how to guide customers to find what they want in a forceful way and promote the selling rate at the same time. In today s competitive environment, direct marketing is an efficient and effective strategy for establishing customer relationship, decreasing marketing cost and gaining customer satisfaction. Furthermore managers are expected to contact with core customers from a limited budget to maintain and satisfy the requirements of customers along with promoting direct marketing systems. Direct marketing system could provide better services via the seamless integration of diverse approaches towards collecting, organizing, storing, accessing, and applying knowledge. Since recommender systems have been increasing gradually, it is difficult for decision makers to find target customers [1]. Predicting the response of customers and making correct decisions is one of the most principal needs of direct marketing. Prediction of target customers has been a common research topic for many years. Any system that can consistently identify target customers in the dynamic market would make the owner of the system very wealthy. Thus, many researchers, marketing experts, and advertisers are constantly looking for this advanced system which will yield accurate responses. Prediction of the customer response is not a simple task. Mainly as a consequence of the close to random-walk behavior of a market data due to the fact that markets have numerous underlying factors, most of which are presently not fully understood, therefore a non-linear model is valuable [2]. A large set of interacting input series is often required to explain a customer behavior. If we are able to predict customer response more accurately, we can avoid wasting resources [3]. There are different approaches to response modeling [4]. Conventional statistical models including moving average, ARIMA, and exponential smoothing are linear in predictions of the customer’s responses. But relationships between target variable and influencing factors are nonlinear and complex. It is hard or even impossible to have a precise mathematical model describing these relationships [5]. Models commonly used in the direct marketing arena to predict response to mailings and other forms of direct marketing promotions are increasingly being used to up-sell or cross-sell customers who contact companies through call centers. The models can be used to decide which of various possible products or services to offer the customer based on a predicted probability of accepting an offer that is estimated on the fly from data already available on the customer or obtained with a couple of questions. A class of such models is called response models, in which the dependent variable is a simple response or not [6]. Response modeling for database marketing is concerned with the task of modeling the