K-均值聚类算法，属于无监督机器学习算法，发现给定数据集的k个簇的算法。

首先，随机确定k个初始点作为质心，然后将数据集中的每个点分配到一个簇中，为每个点找距其最近的质心，

将其分配给该质心对应的簇，更新每一个簇的质心，直到质心不在变化。

K-均值聚类算法一个优点是k是用户自定义的参数，用户并不知道是否好，与此同时，K-均值算法收敛但是聚类效果差，

由于算法收敛到了局部最小值，而非全局最小值。

K-均值聚类算法的一个变形是二分K-均值聚类算法，该算法首先将所有点作为一个簇，然后将该簇一分为二，

之后选择其中一个簇继续进行划分，选择哪一个簇进行划分取决于对其划分是否可以最大程度降低SSE的值。

Yahoo有一个placefinder的API可以用于转换地址和经度纬度。

-K-means clustering algorithm, which belongs to unsupervised machine learning algorithms for a given data set k clusters algorithm found.

First, k is randomly determined as a centroid of the initial point, and then assigning each data point to a cluster set, find a nearest point to each centroid,

It is assigned to the corresponding cluster centroids, update the centroids of each cluster until the centroid not changed.

K-means clustering algorithm k is an advantage of user-defined parameters, the user does not know whether it is good, at the same time, K-Means clustering algorithm converges but poor results,

Since the algorithm converges to a local minimum instead of the global minimum.

A variant K-means clustering algorithm is K-means clustering dichotomy algorithm will first of all points as a cluster, then the cluster into two,

After selecting one of the cluster continues to be divided, to choose which one cluster is divided depends on whether you can reduce the value of t