GA를 이용한 정보서비스 만족도 조사 응답자 분류 예측 모형 개발

The response rate in the survey affects the reliability which may produce the differences in strategy. Therefore, the measure that can increase the response rate in the survey must also be considered. However, the research related to this field has not been conducted enough despite the fact that response rate is less than 30~40%. In this study, to predict whether the respondents will participate in the information satisfaction survey questionnaire, we derived behavioral characteristic metadata of individuals from an operational data. The experimental showed that the respondent classified prediction model # 3 showed the best prediction accuracy of 93.4% with the experimental data, and 93.9% with the validation data. In this study, we think it is noteworthy that we suggested a way for information service users to conduct an effective survey by finding out the factors that can affect the survey questionnaires and able to predict the response probability conditions through GA(Generic Algorithm) which are derived from the business behavioral characteristics of the individual.