Abstract
As one of the most delicate and complex organs, brain contains a wealth of physiological and pathological information. It dominates the human’s thinking and physical activities, and is the highest leading center of controlling emotions and actions. So brain contains a large amount of information, which also leads to the different forms and characteristics of brain electrophysiological signals, making it very complicate to extract and recognize its feature information. Aiming at the characteristic of a large amount of information contained by brain, based on the basic principle of SVM algorithm and K-nearest neighbor algorithm, in order to further improve the discrimination of polygraph technology degree, the present paper puts forward a new lie detection algorithm with K-SUM classifier. And from the biographical information and mock crime, it is confirmed that classification accuracy of K-SUM algorithm is significantly higher than that of single SVM algorithm. This algorithm improves the lie detection ability for subjects, enabling the policemen in the law enforcement process accurately judging whether the suspects have committed a crime, and taking corresponding measures to maintain social order.