现代计算机(普及版)
現代計算機(普及版)
현대계산궤(보급판)
MODERN COMPUTER
2015年
2期
15-19
,共5页
击键特征%身份识别%加权贝叶斯
擊鍵特徵%身份識彆%加權貝葉斯
격건특정%신빈식별%가권패협사
Keystroke Characteristics%Identity Authentication%Weighted Bayesian
生物击键是以人的行为特征为基础的身份认证技术。在朴素贝叶斯分类理论的背景下,提出一种改进的加权贝叶斯方法。实验结果表明,加权后错误率较朴素贝叶斯大大降低,错误拒绝率FRR和错误接受率FAR分别为2.5%和1.4%。
生物擊鍵是以人的行為特徵為基礎的身份認證技術。在樸素貝葉斯分類理論的揹景下,提齣一種改進的加權貝葉斯方法。實驗結果錶明,加權後錯誤率較樸素貝葉斯大大降低,錯誤拒絕率FRR和錯誤接受率FAR分彆為2.5%和1.4%。
생물격건시이인적행위특정위기출적신빈인증기술。재박소패협사분류이론적배경하,제출일충개진적가권패협사방법。실험결과표명,가권후착오솔교박소패협사대대강저,착오거절솔FRR화착오접수솔FAR분별위2.5%화1.4%。
Biological keystroke is an identity authentication technology based on people's behavior characteristics. Under the background of naive Bayesian classification theory, proposes an improved weighted Bayesian method. Experimental results show that the error rate after weight-ed is greatly reduced compared to naive Bayesian method, false rejection rate and false pass rate are 2.5%and 1.4%respectively.