A Method of Online Writeprint Identification Based on Principal Component Analysis

With the rapid development of Internet communication technologies, misuse of online messages for illegal purpose has become increasingly serious. The anonymous nature of online messages makes identity tracing a critical problem. We proposed a model of online write print identification based on principal component analysis for identifying authorship of online-message. The model use principal component analysis algorithm to achieve dimensionality reduction of feature-text matrix, the performance of identification model before and after using principal component analysis were compared by conducting the experiment on online reviews, and then the relationship of feature dimension and identification accuracy was analyzed. The experimental results based on consumers´ reviews of online businesses showed that the model has good performance and efficiency while maintaining the high identification accuracy.