In many practical machine learning and data mining applications, unlabeled training examples are readily available but labeled ones are fairly expensive to obtain. Therefore, semi-supervised learning algorithms have attracted much attention. In this paper, semi-supervised regression via local block coordinate algorithm is proposed. This algorithm preserves more geometrical knowledge of the high-dimensional data by local tangent space alignment, we take the method of automatic alignment of local representations to realize preserving the linear projection between every local coordinate and the g...