中文说明：聚类算法：k―medoids 方法。这儿选取一个对象叫做mediod来代替上面的中心 的作用，这样的一个medoid就标识了这个类。步骤： 1，任意选取K个对象作为medoids（O1,O2,…Oi…Ok）。 以下是循环的： 2，将余下的对象分到各个类中去（根据与medoid最相近的原则）； 3，对于每个类（Oi）中，顺序选取一个Or，计算用Or代替Oi后的消耗―E（Or）。选择E最小的那个Or来代替Oi。这样K个medoids就改变了，下面就再转到2。 4，这样循环直到K个medoids固定下来。 这种算法对于脏数据和异常数据不敏感，但计算量显然要比K均值要大，一般只适合小数据量。 这里是MAtlab源代码。-err
English Description：
Clustering algorithm: k-medoids method. Here we select an object called mediod to replace the role of the center above. Such a medoid identifies the class. Steps: 1. Randomly select k objects as medoids (O1, O2 Oi… Ok）。 The following are cyclic: 2, divide the remaining objects into various classes (according to the principle closest to medoid); 3, for each class (OI), select an or in order, and calculate the consumption after replacing oi with or -- E (or). Choose the or with the smallest e to replace OI. In this way, the K medoids will be changed. Next, we will go to 2. 4. Cycle until K medoids are fixed. This algorithm is not sensitive to dirty data and abnormal data, but the amount of computation is obviously larger than K-means, which is generally suitable for small data. Here is the matlab source code. -err