Voice phishing detection technique based on minimum classification error method incorporating codec parameters

The authors propose an effective voice phishing detection algorithm based on a Gaussian mixture model (GMM) employing the minimum classification error (MCE) technique. The detection of voice phishing is performed based on the GMM using decoding parameters of the 3GPP2 selectable mode vocoder (SMV) codec directly extracted from the decoding process of the transmitted speech information in the mobile phone. The authors´ approach is further improved by the MCE scheme in that different weights are assigned to each likelihood ratio and is considered to be new. The experimental results show that the proposed method is effective in discriminating between true statements and lies.