CLASSIFICATION MODEL SELECTION METHOD FOR DISCRIMINATING FAKE REVIEW

摘要

The present invention relates to a classification model selection method for determining a fake review which comprises: a collecting step of collecting review information; a classification step of classifying the review information into real review information and fake review information; a verification step of verifying a linguistic difference between the real review information and the fake review information; a model constructing step of constructing a classification model capable of determining a fake review based on the linguistic difference; and an accuracy calculation step of calculating the determination accuracy of the real review information and the fake review information by inputting the review information into the classification model. According to the present invention, in a review written in Korean, fake review determination accuracy for each classification model for determining the fake review can be effectively calculated and an optimal model for determining the fake review can be easily selected by using the same. And, by using the optimal classification model selected through the above process, the fake review can be effectively determined.