Identification Model of Commodity False Reviews Based on Integrated Features

Abstract

More and more consumers like online shopping, and false reviews of goods mislead consumers to some extent, so it is necessary to design a false review detection method. Based on the integration characteristics of consumer history reviews, this paper proposes an online false comment detection model. The model introduces time series and combines the static and dynamic features of the reviews to detect false comments. Finally, the model is used to test the comment data of three kinds of Amazon products. The results show that the model has high recognition accuracy.