kNN与kMeans聚类算法的区别

谁借莪１个温暖的怀抱￠ 2022-06-12 04:29 261阅读 0赞

# [kNN与kMeans聚类算法的区别][kNN_kMeans] #

<table style="margin:0px; padding:0px; border-collapse:collapse; border-spacing:0px; border:1px solid silver; word-break:break-word; width:605px"> 
 <tbody style="margin:0px; padding:0px"> 
 <tr style="margin:0px; padding:0px"> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> KNN </td> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> K-Means </td> 
 </tr> 
 <tr style="margin:0px; padding:0px"> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> 目的是为了确定一个点的分类 </td> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> 目的是为了将一系列点集分成k类 </td> 
 </tr> 
 <tr style="margin:0px; padding:0px"> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> KNN是分类算法 </td> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> K-Means是聚类算法 </td> 
 </tr> 
 <tr style="margin:0px; padding:0px"> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> 监督学习，分类目标事先已知 </td> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> 非监督学习，将相似数据归到一起从而得到分类，没有外部分类 </td> 
 </tr> 
 <tr style="margin:0px; padding:0px"> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> 训练数据集有label，已经是完全正确的数据 </td> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> 训练数据集无label，是杂乱无章的，经过聚类后才变得有点顺序，先无序，后有序 </td> 
 </tr> 
 <tr style="margin:0px; padding:0px"> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> 没有明显的前期训练过程，属于memory-based learning </td> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> 有明显的前期训练过程 </td> 
 </tr> 
 <tr style="margin:0px; padding:0px"> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> K的含义：“k”是用来计算的相邻数据数。来了一个样本x，要给它分类，即求出它的y，就从数据集中，在x附近找离它最近的K个数据点，这K个数据点，类别c占的个数最多，就把x的label设为c </td> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> K的含义：“k”是类的数目。K是人工固定好的数字，假设数据集合可以分为K个簇，由于是依靠人工定好，需要一点先验知识 </td> 
 </tr> 
 <tr style="margin:0px; padding:0px"> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> K值确定后每次结果固定 </td> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> K值确定后每次结果可能不同，从 n个数据对象任意选择 k 个对象作为初始聚类中心，随机性对结果影响较大 </td> 
 </tr> 
 <tr style="margin:0px; padding:0px"> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> 时间复杂度：O（n） </td> 
 <td style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> 时间复杂度：O(n*k*t)，t为迭代次数 </td> 
 </tr> 
 <tr style="margin:0px; padding:0px"> 
 <td colspan="2" style="margin:0px; padding:3px; border:1px solid silver; border-collapse:collapse"> 相似点：都包含这样的过程，给定一个点，在数据集中找离它最近的点。即二者都用到了NN(Nears Neighbor)算法，一般用KD树来实现NN。 </td> 
 </tr> 
 </tbody> 
</table>

参考引用：

1.  [http://www.tuicool.com/articles/qamYZv][http_www.tuicool.com_articles_qamYZv]
2.  [http://www.tuicool.com/articles/yUBBbyR][http_www.tuicool.com_articles_yUBBbyR]
3.  [http://stats.stackexchange.com/questions/56500/what-are-the-main-differences-between-k-means-and-k-nearest-neighbours][http_stats.stackexchange.com_questions_56500_what-are-the-main-differences-between-k-means-and-k-nearest-neighbours]
4.  k-Means:http://blog.pluskid.org/?p=17

[kNN_kMeans]: http://www.cnblogs.com/190260995xixi/p/5945652.html
[http_www.tuicool.com_articles_qamYZv]: http://www.tuicool.com/articles/qamYZv
[http_www.tuicool.com_articles_yUBBbyR]: http://www.tuicool.com/articles/yUBBbyR
[http_stats.stackexchange.com_questions_56500_what-are-the-main-differences-between-k-means-and-k-nearest-neighbours]: http://stats.stackexchange.com/questions/56500/what-are-the-main-differences-between-k-means-and-k-nearest-neighbours