Deception Detection Based on SVM for Chinese Text in CMC

Deception is an everyday occurrence across all communication media. Deception detection on Chinese text is vital to the safety of people´s life, the survival of enterprises and the stability of the country. Inundated with massive amounts of textual information transmitted through computer-mediated communication (CMC), people remain largely unsuccessful and inefficient in detecting those messages that may be deceptive. At present it is desirable to propose an automated deception detection method that could help people flag the possible deceptive messages in CMC. In this research we first construct the deceptive and non-deceptive Chinese text corpora, which have not been published so far, and then propose a novel SVM-based deception detection method of Chinese text based on the corpora analysis results. Our method fully combines the deceptive corpora characteristics and the better SVM classification performance, along with feature selection and parameter adjustment, obtaining the three respective deception detection models by using the three different kernel functions. The experimental results show the precision rate, the recall rate and the F- value may achieve 78.3%, 72% and 0.75 separately in open test.