DOI: 10.5716/2251-2136_ICT07
Authors: Cheng-Yueh Tsai, Tzu-Yuan Lee, Chih-Hung Hsu
Abstract:
In marketing area, how to find the latent customers is very important topic. The early strategy of market is STP (segment of market, select target of market and position of market). However, investigating data to predict customer’s need for segmenting market is hard work. The aim of this paper is designed a forecasted-customer system. This system was built by artificial neural network. In the market, this system can help enterprises to find the latent customers. When we inputting some customer’s data into system, like as: sex, age, education, income and career, the system would report the trend of customer’s consumption. After our training and testing, the best accurate identify rate of this system are 84.81{6e6090cdd558c53a8bc18225ef4499fead9160abd3419ad4f137e902b483c465} for trained set and 82.96{6e6090cdd558c53a8bc18225ef4499fead9160abd3419ad4f137e902b483c465} for tested set respectively. The epoch and the number of hidden neurons in hidden layer affected the performance of system. All discussions would in this paper.
Updating...