计算机与数字工程2019，Vol.47Issue(7)：1788-1791,1821,5.DOI:10.3969/j.issn.1672-9722.2019.07.045
基于SVR的人脸记忆分数预测算法研究
Prediction Algorithm of Face Memory Score Based on SVR
摘要
Abstract
A prediction algorithm of face memory score based on SVR is proposed. The algorithm is based on the facial propor?tional features,eyes features,eye-mouth features and HOG features. The face memory scores is predicted based on SVR with multi-features. Analysis of the correlation and error between the predictive face memory score and the real memory score are made with Spielman rank correlation coefficient and residual sum of squares. The experimental results show that the new feature training model is effective. The Spielman rank correlation coefficient is 0.4266,which shows the predictive memory scores is credible. The residual sum of squares Res is 0.0067,which shows the prediction scores is accurately.关键词
支持向量回归/多特征/记忆分数/预测算法Key words
SVR/multi-features/memory score/prediction algorithm分类
信息技术与安全科学引用本文复制引用
唐飞,沈琳琳..基于SVR的人脸记忆分数预测算法研究[J].计算机与数字工程,2019,47(7):1788-1791,1821,5.基金项目
广东省自然科学基金项目(编号:2015A030313587) （编号:2015A030313587）
深圳市科技计划项目(编号:JCYJ20170303094658400)资助. （编号:JCYJ20170303094658400）
