当前位置：
首页 资源下载
搜索资源 - global nearest neighbor
搜索资源列表
-
0下载：
In this paper, we present two novel class-based
weighting methods for the Euclidean nearest neighbor algorithm
and compare them with global weighting methods
considering empirical results on a widely accepted time series
classification benchm
-
-
0下载：
In this paper, we present two novel class-based
weighting methods for the Euclidean nearest neighbor algorithm
and compare them with global weighting methods
considering empirical results on a widely accepted time series
classification benchm
-
-
0下载：
关于metric learning的综述，涉及到许多的知识：SVM、kernel、SDP等-This paper surveys the field of distance
metric learning from a principle perspective, and includes a broad selection of recent work. In particular, distance metric learning is reviewed under different
-
-
0下载：
NW小世界网络的构成原则为：从一个环状的规则网络开始，网络含有N个结点，每个结点向与它最近邻的K个结点连出K条边，并满足N>>K>>In(N)>>1。随后进行随机化加边，以概率p在随机选取的一对节点之间加上一条边。其中，任意两个不同的节点之间至多只能有一条边，并且每一个节点都不能有边与自身相连。改变p值可以实现从最近邻耦合网络(p=0)向全局耦合网络(p=1)转变。在p足够小和N足够大时，NW小世界模型本质上等同于WS小世界模型。
-NW constitu
-
-
0下载：
In the proposed algorithm, these two issues are solved by a novel predicate, which is defined by the sequential probability ratio test (SPRT) and the minimal cost criterion. Starting from an over-segmented image, neighboring regions are progressivel
-
