Neural network applications for buyers' intelligent decisions in bookstore business

The Artificial Neural Networks (ANN) approach performs best when we solve real world problems which are fallen into the following 4 categories. 1:Pattern Classification such as quality control, financial forecasting, laboratory research, targeted marketing, bankruptcy prediction, optical character recognition. 2: Associative Memories, 3: Feature Extraction, and 4: Dynamic Networks such as speech recognition, adaptive control, time series prediction. financial forecasting, radar/sonar signature recognition and nonlinear dynamic modeling. ANN solves these problems like human beings do, i.e.. they observe events to extract patterns, and then make generalizations based on their observations. In this paper we challenge to apply ANN to bookstore buyers' "where-to-buy" decision making process which is one of the Pattern Classification problems. Randomly chosen real purchasing data of buyers are used to train the ANN. We analyze the ANN's decisions of how reasonably the ANN is trained and how much the ANN is close to actual buyers' intelligent decisions. As a result, we show not only the usefulness of ANN-based approach in such business but also potential applications to any business involved in complicated decision making processes.