前向网络利用反向传播算法训练多层网络，使训练后的网络较好地拟合样本空间中各点的函数值。
Feedforward networks use back propagation algorithm to train a multi-layer network. After training, the multi-layer network can fit the function in the data space very well.
基于图的学习是近几年来半监督学习中一个相当活跃的方向，它用图来描述样本空间，利用近邻点的位置来控制标记信息的传播。
Graph-based learning is a very active direction of semi-supervised learning in recent years. It describes the sample space by graph, and USES neighbors to spread label information in point cloud.
基于这一问题本文通过构造统计量对所给的样本点进行选择，剔除对模型的构造有很大影响力的样本，从而获得一个相对合理的样本空间。
Based on this problem, this article selects the sample points by constructing statistics. First, it removes the outliers to have a relatively reasonable sample space.
应用推荐