poweriterationclustering-文庫吧資料
2025-07-23 16:19本頁面
【正文】 n: ? A subsequence of converges to an eigenvector associated with the dominant eigenvalue kbPower Iteration Clustering(PIC) Unfortunately, since the sum of each row of NA is 1, the largest eigenvector of NA (the smallest of L) is a constant vector with eigenvalue 1. Fortunately, the intermediate vectors during the convergence process are interesting. 222| | | | a n d s ( x , x ) = e x p ( )2iji i j xxxR ????Example: Conclusion: PI first converges locally within a cluster. PI’s Convergence Let: W = NA (Normalized affinity matrix ), 1 1 n112W h a s e ige nv e c tor s e , ... , w ith e ige nv a l ue s λ ,..., λλ = 1, e is c on sta nt, ... ., a r e l a ge r tha n the r e m a ining o ne snke??Spectral representation of Spectral distance between a and b: 122 h a s a n ( , ) e i g e n g a p b e t w e e n t h e k a n d ( k + 1 ) e i g e n v e c t o a n dr t h t kkhW ???? ???? ??? ?e v e r y W is e bo un de d? ? i s t h e t t h i t e r a t i o n o f P Itv0( , ) d ista n c e b e tw e e n a a n d b : tv1 2[ ( ) ( ) ]tnjj j jjke a e b c ???????? ?????? ?2si gna l is a n a ppr oxim a t i on of spe c , buta ) c om pr e sse d to the sm a l l r a
研究報(bào)告相關(guān)推薦
鄂ICP備17016276號(hào)-1