[english][all] (叫猔種いゅセゼ繦璣ゅセ˙穝)
Slides
顶糷Αだ竤猭hierarchical clustering硓筁贺顶糷琜篶よΑ盢戈糷糷は滦秈︽だ吊┪籈玻ネ程攫挡篶盽ǎよΑΤㄢ贺
セ竊盢皐癸籈Αパτ顶糷だ竤猭ㄓ秈︽弧
- 狦蹦ノ籈よΑ顶糷Αだ竤猭パ攫挡篶┏场秨﹍盢戈┪竤籈硋Ωㄖ
- 狦蹦ノだ吊よΑ玥パ攫挡篶郴狠秨﹍盢竤籈硋Ωだ吊
籈Α顶糷だ竤猭agglomerative hierarchical clusteringパ攫挡篶┏场秨﹍糷糷籈秨﹍и盢–掸戈跌竤籈cluster安砞и瞷局Τn掸戈玥盢硂n掸戈跌n竤籈ョ–竤籈掸戈
絛ㄒ甶ボ籈Α顶糷だ竤猭┮玻ネ攫瓜dendrogram
- 盢–掸戈跌竤籈 Ci, i=1 1 to n.
- т┮Τ竤籈丁禯瞒程钡ㄢ竤籈 CiCj
- ㄖ Ci Cj Θ穝竤籈
- 安ヘ玡竤籈计ヘи箇戳竤籈计ヘ玥は滦狡˙艼竤籈计ヘ盢и┮璶―计ヘ
˙艼い孔禯瞒程钡ㄢ竤籈 CiCj㎡ㄆ龟﹚竡ㄢ竤籈ぇ丁禯瞒Τ贺ぃよΑ–贺よΑ┮眔挡狦常ぃび硂ㄇ盽ノ竤籈禯瞒﹚竡弧
- 虫硈挡籈簍衡猭single-linkage agglomerative algorithm竤籈籔竤籈丁禯瞒﹚竡ぃ竤籈い程钡ㄢ翴丁禯瞒 $$d(C_i, C_j)=\min_{\mathbf{a}\in C_i, \mathbf{b}\in C_j} d(\mathbf{a}, \mathbf{b})$$
- Ч俱硈挡籈簍衡猭complete-linkage agglomerative algorithm竤籈丁禯瞒﹚竡ぃ竤籈い程环ㄢ翴丁禯瞒 $$d(C_i, C_j)=\max_{\mathbf{a}\in C_i, \mathbf{b}\in C_j} d(\mathbf{a}, \mathbf{b})$$
- キА硈挡籈簍衡猭average-linkage agglomerative algorithm竤籈丁禯瞒玥﹚竡ぃ竤籈丁翴籔翴丁禯瞒羆㎝キАㄤい ボ 戈计 $$d(C_i, C_j)=\sum_{\mathbf{a}\in C_i, \mathbf{b}\in C_j} \frac{d(\mathbf{a}, \mathbf{b})}{|C_i||C_j|},$$ where $|C_i|$ and $|C_j|$ are the sizes for $C_i$ and $C_j$, respectively.
- ║紈猭Ward's method竤籈丁禯瞒﹚竡盢ㄢ竤ㄖ翴ㄖ竤いみ禯瞒キよ㎝ㄤいm ボ Ci″Cj キА $$d(C_i, C_j)=\sum_{\mathbf{a}\in C_i \cup C_j} \|\mathbf{a}-\mathbf{\mu}\|,$$ where $\mathbf{\mu}$ is the mean vector of $C_i \cup C_j$.
иㄏノぃ竤籈禯瞒ㄓ玻ネ顶糷Α竤籈攫┮眔挡狦
パ瓃竤籈攫и祇谋疭┦
- single linkage 穦竤籈筁祘い玻ネ狦
- τ complete linkage ㎝ average linkage ゑ耕甧玻ネ霍繷秈狦
璝璶芠竤籈筁祘い笆篈甶ボ叫ǎ絛ㄒ絛ㄒ single linkage 竤籈禯瞒
ㄆ龟и靡狦戈栋琌蝴キ翴┮Θ栋τи蹦ノ single linkage玥┮眔硈挡瓜琌硂ㄇ翴 minimum spanning tree
俱砰ㄓ弧顶糷Αだ竤猭纔翴
琌顶糷Αだ竤猭Τ翴ウ硄盽続ノぶ秖戈螟矪瞶秖戈
- 阀├虏虫ノ攫挡篶ㄓ瞷俱璸衡筁祘
- 惠璶戈翴ㄢㄢぇ丁禯瞒碞篶だ竤挡狦τぃ惠璶戈翴龟悔畒夹–戈翴ボンτぃゲ琌丁い翴
Data Clustering and Pattern Recognition (戈だ竤籔妓Α侩粄)