首先是在“人脸流形”上的识别,随后应用LE的线性化算法LPP进行人脸识别,解决了流形学习对样本外(Out-of-Sample)的学习问题。应用UDP算法进行人脸识别实验,实验结果表明UDP算法较好的解决了流形学习的分类问题。
基于17个网页-相关网页
发现宏观经济变量的样本外预测能力仍然相当的显著。
The results show that macro economic variables have obvious out-of-sample explain ability.
文章从样本内拟合能力、样本外预测能力、利率曲线形态方面比较了各个模型的表现。
We compared the models from performance of in-sample, performance of out of sample, and the shape and robustness of interest rate curves.
模型的跨度为一年的样本外条件异方差预测,显示出该年末汇率的震荡,与实际情况一致。
Out-of-sample volatility prediction performance of one year confirms the actual higher volatility in the end of the year.
应用推荐