Online Survey Prediction Model for High Response Rate via Decision Tree

In recent years, online survey has received many attractions and become an important way for companies to promote their business development. According to the characteristics of the accumulated online survey data, we present a new approach to construct a prediction model which is based on ID3 algorithm to forecast users' response behavior. Especially, we introduce XML to store the prediction model on the basis of the structure property of decision tree. We collect lots of online survey data and validate the effectiveness of our proposed approach. According to analysis of experiment results, the main contributions of this paper are three-fold: (1) response rate based on prediction model of decision tree is higher than that of random way, (2) the improved prediction model of ID3 can improve the crawling rate of classic ID3 algorithm, (3) the cost of online survey can be reduced by the prediction model effectively.