基于密度聚类的DBSCAN和kmeans算法比较
根据各行业特性,人们提出了多种聚类算法,简单分为:基于层次、划分、密度、图论、网格和模型的几大类。
其中,基于密度的聚类算法以DBSCAN最具有代表性。
场景 一
假设有如下图的一组数据, 生成数据的R代码如下
x1 <- seq(,pi,length.out=)
y1 <- sin(x1) + 0.1*rnorm()
x2 <- 1.5+ seq(,pi,length.out=)
y2 <- cos(x2) + 0.1*rnorm()
data <- data.frame(c(x1,x2),c(y1,y2))
names(data) <- c('x','y')
qplot(data$x, data$y)
aaarticlea/png;base64,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" alt="" />
用密度聚类DBSCAN方法,可以看到聚类效果如下:
library(ggplot2)
p <- ggplot(data,aes(x,y))
library('fpc')
model2 <- dbscan(data,eps=0.6,MinPts=)
p + geom_point(size=2.5, aes(colour=factor(model2$cluster)))+theme(legend.position='top')
aaarticlea/png;base64,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" alt="" />
同样,请读者看下k-means的聚类效果。
# 用K均值聚类
model1 <- kmeans(data,centers=,nstart=)
library(ggplot2)
p <- ggplot(data,aes(x,y))
p + geom_point(size=2.5,aes(colour=factor(model1$cluster)))+theme(legend.position='top')
aaarticlea/png;base64,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" alt="" />
场景 二
假设有如下一组数据,生成数据的R代码如下。
set.seed()
n <-
data <- data.frame(cbind(runif(, , )+rnorm(n, sd=0.2), runif(, , )+rnorm(n,sd=0.2)))
names(data) <- c('x','y')
library(ggplot2)
qplot(data$x, data$y)
aaarticlea/png;base64,iVBORw0KGgoAAAANSUhEUgAAA8IAAAKfCAMAAACSUoECAAAAt1BMVEUAAAAAADoAAGYAOpAAZrY6AAA6ADo6AGY6kNtmAABmADpmOpBmZmZmtv9/f39/f5V/f6t/lcF/q9aQOgCQkGaQ2/+Vf3+Vf5WVf6uVlX+VlcGVq6uVweurf3+rf5Wrf6ur1v+2ZgC225C2/7a2///BlX/BlZXB6//Wq3/W///bkDrb25Db/7bb/9vb///l5eXrwZXr1tbr///y8vL/tmb/1qv/25D/68H//7b//9b//9v//+v///+zk6n+AAAgAElEQVR4nO2dC5vjNLpuzWXOgZlDwcBmbhsY+myYpnvPIb3pbpqi8/9/16nkS2zJlh1f9CqStdYDXYnjqvjNpxXJsp00RwAomObeGwAAW0BhgKJBYYCiQWGAokFhgKJBYYCiQWGAokFhgKJBYYCiQWGAokFhgKJBYYCiQWGAokFhgKJBYYCiQWGAokFhgKJBYYCiiaDwYRWv1/3aNlZu6yYIqqSyoChcWcHTQlAlKGxUVvC0EFQJChuVFTwtBFWCwkZlBU8LQZWgsFFZwdNCUCUobFRW8LQQVAkKG5UVPC0EVYLCRmUFTwtBlaCwUVnB00JQJShsVFbwtBBUCQoblRU8LQRVgsJGZQVPC0GVoLBRWcHTQlAlKGxUVvC0EFQJChuVFTwtBFWCwkZlBU8LQZWgsFFZwdNCUCUobFRW8LQQVAkKG5UVPC0EVYLCRmUFTwtBlaCwUVnB00JQJShsVFbwtBBUCQoblRU8LQRVgsJGZQWfR9PEec7sg8YChQ0KrmRB0KaJ5HDuQaOBwgYFV9ILOiFp08RyOIegSUBhg4Ir8YNOSYrCiylQ4Xd/e3E8vv/x4YsXdt+5icI3uX/QaUsZSC+lPIUfvz0J+9OXp//OODdR+Cb3D3qjo2U6ayHFKfzq4T+feuHH755bb3x0b6LwbTII2hkcS9cQGQRNQ3EK/88vJ1/fXTQ+Ht2bKHybHIJ2BgsdziFoEopT2JwNKvxwZsZfgDvSNN2tE/fcFojPJoWNde8qvGcrcYJaz3vuff2d4tgd8r2DJmNPvTAKz+DOQZuWQzeQNqkjO0xFlWxXmOmstWSjsDnsL4v5nFRUSYTjwhxUWsm9gw59bZpGIfG9gyajVIUfvz2fz/HTN+1NFJ7F3YP2R81No3H47kFTUaDCN1i3SRRcySDoqMHxHM4jaAJQ2KDgUemZOB7UdReFV4HCBgWPSV/FOQoHfm0TVFQJCht7LfigOx0Nel3TOuGu8+4ONq2GiipBYWOvBV+qcHNo/F9pvd7gMBVVgsLGbgs+YyDdPyTc9cf+0tXbQEWVoLCx34LfnM7q+lmnE77cQuFFoLBBwZUEDip5mvYnpV3B1z5nHkETgMIGBVcyELE/gHamtdrh9EaHqagSFDZqKfhQxBGF2x/eauuelIoqQWGjkoIHRBwZSLvDaGe1Vc9KRZWgsFFJwV0Pe3JeutzgoNpZrftL85+ViipBYaOWgvtKutf731LYd3hRj0xFlaCwUU3Bh32veyfk8CEwpbVsVE1FlaCwUVnB+xcLTzh8cD/SwzMYhfugsEHBlbRB+w4fPGNdSb21ektmQUWVoLBRWcFPDK4p9Lvk6wmWo9PY85+TiipBYaOygl/xVfTnqiZPkV50eCmDoGlAYYOCKwl/FHzg0sLQtNZKcgiaBBQ2nNfBazfKrwypreAu7hkcnaoRDc4lqB4UNrrXwWs5G5vRDSoruEN/6vkyho5ocCZBE4DCRvs6eG1ne0OapLKCd3iSXm50/8R52bMImgIUNlBYyajC7b3+QaXtr3oWQVOAwgYDaSXBgbQ3nRVr+NyRR9AEoLDBdJaS4HRW774zkI5CJkH1oLAx43WIb3NlBZ+k3TGO9Zy5Bo0OChtjr0PXrgRj6soKPk3sVzfboLFBYWPkdfDnSiP3E7UVfD4xXuMigsYAhU80zWt3esVZPjhQGbMzrqngS161KK8xCivJTeHRyVBX3oiHLK9UVPAlr1qc1xiFlWSm8MQBDXea1DuVqH189YtwqKngi6xE4UWg8PSXBwx2f2MeNq6n4MusZCC9BBQ+TAyk+8J2/zq/t+IJL1RU8GUvFNNZC0DhE9PTWc6diYdXUFPBpefIhEBhJdkpPPY6uI72fXWmp9c2z8oKnhaCKilG4UOvE3Ycvk5QHzbsulVW8LQQVEk5Cg/3fAefzrZhOF1ZwdNCUCWlKBzwtx06o/AC3KCpdorvHjQV1Srcm2kOvQ7DUfTVXH9QzUD6Fk7QKAeM5nDvoMmoVeG+gf3XoWmGY+XG+/7bCJ8LUFnBT2ydxJ8PCiu5v8IDPV+HHr/o6u4GOwPpAwPpuaCwEBQOKeyPm71To9vfcm1mID0NA2khlSp8YyDt7/K2vW//RMvBmHrV65CWu7dsprMiU6vCN6azvB6264D9w8L+o6tfh7TQspVUFvS+CvcITGcd/IsNA6dqbeyEayt4WgiqpACFL7gOBxXetG9XWcHTQlAl5SjsDaUDDh827dtVVvC0EFRJQQp7p1UOHF73XC2VFTwtBFWSocKjOrpTVv6O7/ap1coKnhaCKslP4WCX6k9pja62nsoKnhaCKslO4eC8srsDfN3rdVYbdX4+lRU8LQRVUoTCgyO/7bR0+2jgr6x4HdJCy1ZSWdCMFA7Z1xPYv+8s7/3C8tchLRm0bM6RjgoKG6Ofm9Vq2QQI/cLi1yEt92/ZkScUxhj7KCXpk1ZW0awUHr4OVyNHrhlmIL0A/xzpNA6PfIiD9qkrq2gJCg+XjBvMdNYouSg8/dwRtqqyiuatsF9t95jSZXw9+BPLW0BlBb9wx4H0pMIxtquyihakcDel1f17feTQW2cBlRX8yt2ms24azPl2cylBYXcg7e0W967876+z/HVISwYKJyE4TJoeRqPwfEpQ+Lq/61zf0P1zaJwf6z9+Z7cF778Q6YN6xRie4X7zV1ay24r2KUBhb9rK+bcZPuB0w6teh7SkKPjgpUgbtH8Z9+XmzTdZprPmk7/CQ11DCnsnXNY2nbVkSJo0qFMef3MGxYq/U150RZcgVfh1DC5t4HXXHBpb+Lq7ebmx4VnibOt9sPijD/kPpgzqVqy3Od1GOaWMSskVXYQFFSm87l0lNJBuf3Z3Dt6O1cp3sJaC37O9fi70oLcgZdB2fDS+Oe6AKm5XXHBFl5H/QLobZI2OuOrec2puOOzfTz6QPnhBJ89/j/BW3FFwRZdRgsIpKLng/eY/KULy6azDdNBG5XDJFV0EChtFF9xt/ZMiPD2QWdC+wfFG05kF1YHCRuEF7xk8osHpgbyC9uTtHVjYRF5BheSvcJpzAEsveP8yzPA6sfc3Z2zRiVkKH4ZHkTdRekVnk73CTkWVjW8nBb/ZCSeT2H2a0U8GH24SCi8md4V7e3nbwk6wl4L3FPBfsYQOe08zdsl/E9ggBtJLKUdhbdvbTcG9l6j/iqVz+LbCnsHOQIvprIXkrnB/omZT2AmKLPitl6P/knWXg6Tphq+3bysc+KWtFFnRNWSv8EFS3wG5F3zsgr0bv9QNYS53D8d0O8PtzWmF25PsRt6lV25t7hWNRv4Kd1Q8nRVo2qHuNHj24vXXbf1jojl+90leh0+p8xKMjvHXvuNkXtF4lKSwkrwLHrR1uNDt1Jx1vDFroqDepoUt9JydNHiVw3lXNCIobORd8HGFA4sCHZvzSJqg3saNDReGCgcGCCh8CxQ2Mi/4WPfU3Tl4yh4OYYNT9cI3FG7v++83bZLgugvJvKLxQGEj94IHp7OcmwGHwwonOzdrYiDtbEj/4ozge9WqLci9otFAYaPogoesPXiTWMkVnpzOGt+UiJtYdEWXgMJG0QX3ejVHj95uZsKBtM/YBxqOHChD4UWgsFF0wa8N3zl1wz9H6/oj2XSWT/9knf4BpXZ5zDNP7h80EShsFF1wp58N2OErcUw2lO5wg46O6q/vQoOtX/mkRVd0CShslF1wR96Bw5NCp8H74J0JgwMOr9/csiu6ABQ2Si54431Kfvi0p27dOzjcBR01eDAl11u6nJIruggUNgoueNvIewa7e8feyvdSeDg/7q7ke+ssROFpUNgot+Be9xU4GuxOZx3uOZA+P3Nv29xt6R8i7n5pFeVWdCEobJRb8MH4eeCw5+7dprPCW3Z58xn/XaazboHCRsEFdzrhcY1b0+8WNDA28EXu8kR50oIrugwUNkouuLsr7PZv3kdTSRW+YV03kO4MHiwJ7AlvouSKLgKFjZIL3jX+Q6+zC0wfKYLesu51t5qncGh2y3t8CyVXdBEobBRc8ME41HOir4gg6E3rXnsruusOHUbhpaCwUW7BB22+19c5vfHp7j0Vvm7LcFudv2E/t1tcbkUXgsJGuQV3+i9nSXcMqRtfu/ejMnMgfRia6fe+3ToReuJyK7oQFDYKLvjAAe8Iq7ejHPEqgt4mTD488WlorsP95du2qeCKLgOFjZIL7hwVHjziLxzujCbhOiPdX+7vBQ8e2PakJVd0EShsFF7wETV7C8eMUdMeF+4tH1WYgfQCUNgou+CjZmZh8DnoxFtM+JGtT1p2RReAwkbZBR83MziMvqPCw0sbVJtUdkUXgMJG4QWfkqCdih7vrMV0A+nu2Xundcd/0sIrOh8UNkov+JTB3VGaprnjp3YMj/5KhwWlV3Q2KGzstuBNr6u7a9B2M9phte4NZbcV7YPCxk4L3jRZKeycYdIOC1TstKJDUNjYZ8E7d9v+7s5BY1+PNM4+KxoAhY1dFtztfa/G5BJUvkueS1A5KGzssuChqaJdBg1RWVAU3mfBA8PVfQYNUFlQFN5pwf2LHE7sNOiQyoKi8L4L7vTG+w7qUFlQFN51wd194l0HvSD7hKFboLCBwrGpTGELW0FQA4WNfRe8qoH05Q1r/0EvoLCx84LXNJ01Q2HVQWkUNlBYSQVBbwykhVdrobBBy1ZSQ9CR6Sy7XNk5ZS26yChs0LKV1Bu06SE4TxuFjTwKngCCKukH7RvsnD8eDRQ2sih4CgiqpA3qXaKMwigcEYIquQZ19nyv4+d5n8a7Qm8UNmjZSmoL6s1edc7ednhNF43CBi1bST1Bve+v6H3fxe2P7Vo1zEZhg5atpJqgVwd7A2n/YRRG4S0QVEHTfaHjZde3CX5e162jSgykUfgmBBXgC+v1pV2v3K08+ZcWPzkKG7RsJTsP2jSTDvcGyJzagcKbIGh8OkOPQ4X9FVbt7U6DwgYtW8mugk59Edzx0PsCCWcF1bfTFKrw24cT35xvv/vLw8Off0HheRB0I0MDnS9eG5zacfAW3JqRXkWhCp81/uKF/fzSW7xuk2jZSnYUdKige/8YWG24l8xA2nj89nu78dM3KDwfgm5jYj/30A6k+3u/h+BEVyTKVfjVZfD8/l/PUXg+BN1IqBN2FfaUHZ4tPRyEb6VYhd//eOmEH7/7+8PnZ4vP+8cP898EAFbQNP377hLztLvR3ndujf3uLpif593fXlxu/PX58fGfL9oH1r2r0Dkp2XdQpyc9ho8RH7r73i9EGVcX2wu/ciexHr/rBtPrNomWraSmoD1lewo7Pw+1K+xNYqHwbAiqpJ3OOhwuMjd9c71d4hiTW6Uq3O4Kn48tvfsHx4VnQlAlflB3Xsuf3op5tmWpCl/63VNf/Orh4YtuVxiFpyGoEi+oZ2xA4UhHl0pVeJx1m3T/gieCoCIGH0Lr7As71/53S2IMok+gsEHLVrLToK6CJuTRX9CbzvKu+Y8ylXUChQ1atpJ9BnUVvAh5PPgLeru9/V4ZhVF4OQSNRc9PX+GQwYNemYE0Cq+AoLHwu9F2IN0/r8NbezCyjgAKG7RsJfsM6nejl+ms68LrDq+78sDrKKCwQctWstOgQxWPnaHDCxpCe8cRQGGDlq2knqAT4+SBwkxnofAaCBoZT8NuIB0SdDidFcNhFDZo2Up2HNTXsJ3OGkx2OQtjXuNwQOErtGwl+w3a87B/XNi50y099JduA4UNWraS/QYdVXh40eHljvfLMTYBhQ1atpIdBx0MpJ1H2hWiH0hyQWGDlq1kz0F701nDh8UGo/AFWraSuoL2r37QGozCF2jZSqoK2p+xkvp7QOErtGwlNQUdHjfSGozCF2jZSmoKqjj0OwkKG7RsJVUFFRz6nQSFDVq2krqCDq5f0oLCBi1bCUGFoLBBwZUQVAgKGxRcCUGFoLBBwZUQ9Ipg3xiFjTwLLoCgSm4FVcxQo7CRZcEVEFTJjaCS48QobORYcAkEVYLCKCyHoEoYSKOwHIIqYToLheUQVEllQVG4soKnhaBKUNiorOBpIagSFDYqK3haCKoEhY3KCp4WgipBYaOygqeFoEpQ2Kis4GkhqBIUNioreFoIqgSFjcoKnhaCKkFho7KCp4WgSlDYqKzgaSGoEhQ2Kit4WgiqBIWNygqeFoIqQWGjsoKnhaBKUNiorOBpIagSFDYqK3haCKoEhY3KCp4WgipBYaOygqeFoEpQ2Kis4GkhqBIUNioreFoIqgSFjcoKnhaCKkFho7KCp4WgSlDYqKzgaSGoEhQ2Kit4WgiqBIWNygqeFoIqQWGjsoKnhaBKUNiorOBpIagSFDYqK3haCKoEhY3KCp4WgipBYaOygqeFoEpQ2Kis4GkhqBIUNioreFoIqgSFjcoKnhaCKpEq/LocStrWTRB0b1hQkcLr3lV4z1ZCUCH764XXbRIFV0JQIShsUHAlBBWCwgYFV0JQIShsUHAlBBWCwgYFV0JQIShsUHAlBBWCwgYFV0JQIShsUHAlBBWCwgYFV0JQIShsUHAlBBWCwgYFV0JQIShsUHAlBBWCwgYFV0JQIShsUHAlBBWCwgYFV0JQIShsUHAlBBWCwgYFV0JQIShsUHAlBBWCwgYFV0JQIShsUHAlBBWCwgYFV0JQIShsUHAlBBWCwgYFV0JQIShsUHAlBBWCwgYFV0JQIShsUHAlBBWCwgYFV0JQIShsUHAlBBWCwgYFV0JQIShsUHAlBBWCwgYFV0JQIShsUHAlBBWCwgYFV0JQIShsUHAlBBWCwgYFV0JQIShsUHAlBBWCwgYFV0JQIfkr/NunzccoHA2CKqks6EyFj8eXTdN8gsJxIKiSyoLOVniZxes2iYIrIaiQQhQ2iT/8AYW3QlAllQVdoPCbJ38/O/7+1Uc/o/BGCKqksqBzFf7t06Yxd9/M6YbXbRIFV0JQIfkr/NunH3x9W1wUnglBlVQWdKbCS1m3SRRcCUGFZK/ws9MY+uV5XxiFI0BQJZUFnaXwy9Mo+s3TP7/+YabD6zaJgishqJDMFf7t04+Px9+/Op2e9WzOdDQK34KgSioLOkfhZ43LrJM71m0SBVdCUCGZK/zbp5/YOJpeOBIEVVJZ0DkKH599+IONo99wgmUUCKqksqCzFD4NpW1Oeu6FDus2iYIrIaiQ7BVezLpNouBKCCoEhQ0KroSgQgpQ+HSOdDP3OiUUvgFBlVQWdK7Czz76+eXHnNoRCYIqqSzoTIVPx5XefPTz8SUHlWJAUCWVBZ2t8GfHX//3D+f/UXgzBFVSWdCZCv/+1SfH3/74NQrHgaBKKgs6U+Hjyw9/OD77hIF0HAiqpLKgcxU+Pvv4NCs9c0IahachqJLKgs5WeBnrNomCKyGokPwVPu0Hn5j1wVkofAuCKqks6EKFX6JwDAiqpLKgsxR+NnK18Lu/PDz8+Zfzzfc/PnzxAoVnQlAllQVd2Av7vP2yvfnTl6f/UHgeBFVSWdCZCof56Zvrrcfvnh/f/a3rhtdtEgVXQlAhBSgcuszh/b+eX2+e9D1pfDw+nJn7JgAA8Zl7mcPjd39/+Nws7hQ21r2r8J6thKBC8u+Fg5c5vPvr8+PjP1+g8GIIqqSyoLMVHrnM4eItCi+CoEoqCzpT4dHLHC7eMp21CIIqqSzoTIWDlzm8/eLF8d0/7LgwB5WWQFAllQWdq3DwModXD+fzOU7Hlh6/5dSO+RBUSWVBZyu8jHWbRMGVEFQIChsUXAlBhWSu8PW0Dj7BMhYEVVJZ0DkKn7CvYrEvVkLhrRBUSWVBZypsX6l05IN34kBQJZUFnanw6dSOE1zyHwWCKqks6EyFT6d2nODLSaNAUCWVBZ2p8NO+8Kkbfsm+cBQIqqSyoHMVtmnpmX0wCt+AoEoqCzpb4WWs2yQKroSgQlDYoOBKCCoEhQ0KroSgQlDYoOBKCCoEhQ0KroSgQlDYoOBKCCoEhQ0KroSgQlDYoOBKCCoEhQ0KroSgQlDYoOBKCCoEhQ0KroSgQlDYoOBKCCoEhQ0KroSgQlDYoOBKCCoEhQ0KroSgQhYEbZo5i+aAwkbmBY8HQYVMB3UVPX0O7ODhdQ6jsJFfwUUUEnRlj+SQXVBXUfss58HDq1KjsJFdwVWUEXRtj+T9iY1/YAVTQV1FGxSOTxktOwJFBF3QnIer2ZIIbwLLmalwyGAG0lspomXHoIig4TY+tmZoyeo+bRPzBtJxDUbhC0W07BiUEXSugcP1LksyVPjgD6MHD67dXBQ2ymjZESgk6IJOOKhwfgNphxGD6YU3UUjL3k4pQWfvCocH0vlNZ7mEt9reeRY/KQobpbTszewt6Nh0Vs5BQ13u+rEDChsZFzwuBBWyQeHD6kk4FDYyLnhcCCpkqcJjw+lFoLCRccHjUnrQtoXfaupZBB3ZyNbgsR35JaCwkUXBU1B4UO/g6uSaOQQdGS9f/g3vEi9+UhQ2cih4EsoO6gxBbzmcQdDARjpvQbGOXaOwkUHB01BsUN/cwRlOAxcyCDp+1Pq6/TGOfKGwkUHB05B/0HCz7u8+9gwY2pBD0LDBjWNwe8JW+A/OURyFjRwKnoTsg4a6JqfFO9NZngKDX7tfUG8T3SXdJvcUHuuOZ3XTKGxk37JjkXvQ8BxPv9fqrdl/7HTzbkF7W3XwRhC9LXYfWHuMCYWN3Ft2NHIPGmi2jYe/PPhL5zv3CtptS2em/9PZZGd5M+xzUXgJubfsaGQfdLoT9h4b2RW29e6v8KD39ba/dXYi3hyDUfhC9i07FvkHDe8Kdz+Drdp94L4Ku28noRG1f1y4b3B4R2ESFDbyb9mRKDKoO3E72q67B+46kPYHB+32uCNmV9vJ3YRZx5xQ2CiyZa+h9KB+yx4Mn7vFGQR1Rw9j1o5N1qHwUjIoeBoKC+ofOO2f0uTb7Df5+wfttqjX0facDRo8/3xpFDbuX/BElBXUbeWXf3sGh30+5BDU2boxaQ/OzsF1/9j5/VlPisLG/QueiDyDTuzeBqZ8emdHNME/kkHQ4SChOfQdvgZ01lr4pChsZFDwNGQZdKzhjircmwUK/W4OQd13F9fYwN7xZbV+HA4qzSaHgichx6DjHrojz5DJExLnFLTVNPT+4yQYhJnTKaOwkVPBpeQYdFzhwXTWfIczCtp3NJyjfTz0ixOgsJFRwbVkGTTcUMPLphzOa0a6xTHxciNk8vXx8C+Og8JGRgXXkmfQMVvDS8cU9n4jp6DOhnmqXjd9VNQZBqPwhZwKLqWYoGM90LTB3W9kFXRMxOsmTzh880lR2Miq4EqKCeoKGRhdXnov55BqzgobQx8dgxcfTLqCwkaGBddQTtBmrH17O4/9cyzbO/kFHWo6GnEJKGzkV3ARJQQdntI0dHhk97m7ncNlDh4BT8N78QtBYaOElh2FAoJe2/OUwv1GPzQgh4sNPUJdbbtkvcEofKGAlh2H/IP6o8uAwyOLen8mg0v+e3RvTe6yzU+Kwkb+LTsS+Qft7SAOhA1o3f1K92fyUzgw67ZlAH1FqvDrcihpWzdRQNBTu3Z+Pv1jrd55uMVd8tpZ6Y5Bva0IP+wmnFx5DhZUpPC6dxU6JyUlBB1ONgdGzV1ndu2R/f4vu+msC+4IYqLHng8DaaOElh2FQoMG5q5ck3tSn8gjaEBQdycg+4H0uk2iZSvZUdCewTkeFw4q6u7HM50VjSwKnoJ9Be1NbOWm8OhAOcYA+goKGzkUPAn7CjpUOKuB9LTCkRxGYSOHgich/6CLmvbA4Pa3swg6JioKxyeLgqcg+6AL27Z7ODi6wkstmzWddVnOQDoy2bfsWOQedFP/FHkgvXhLFgSNZjAKX8i9ZUcj96DtMd5VzxR1Omv5u8kdK4rCubfsaGQftD1LY+OTzgk65ySMJc+JwgYtW0n+QQfnWa0bdc4IestQ5UA6Hihs5N+yI1FE0J7Cq4bWt4Pe7mW3T2clAIWNIlp2DMoIOjB4+STXQoWjTC+hsEHLVlJIUH8YvcbhZQPpOAd5UNigZSspKah/vDe+wgfP4Mu1Q2OrzGHZ1PuCladAYaOklr2JgoJ20q6Zo14U9Krw4FmWPe2i48KcnRWXglr2NooJ2u94409n9Z9t9MMyZ/+V+UEjnmGJwkYxLXsrpQRdNYXlMjeoe9UfCqPwIgg6zkaDn35xZtDh3Pf4wzdhIG3QspUUEnSrwU0zL6g3jxU6ytT/VL1JmM4yaNlKSgm6pZdaoL83j+X6vOqAEweVDFq2kmKCbuillvTg7jyW63P3Jxb8NRQ2aNlKqgg6MZAezHE781gXWT2dUXgFtGwldQQdnc4aGx937no6L/yYSRQ2aNlK6g46Pj4euOs5PPM5UdjIp+BiCKqkH9TdybX73dzVdQ2v9/V2iOeBwkYWBU8BQZX0gg6GxwfHVGeNwQLRceF4oLBRa8tOQgZBmzGFXXHHuuXZz4nCRgYFTwNBlUwp7JxM6e/59n1d6DAKGxkUPA0EVRIYSHuHiwbuTjg89zlR2Mih4EkgqJLQdJY77+ztC48qvOwsMRQ2sih4CgiqZDRocJ+4E3m4+vznRGEjr4ILIaiSiaADg6+frrekxw2BwkZuBZdBUCVTQV2H/dHzonHzABQ2siu4CoIqmQw66Ia95cO1Zz4nChv5FVwEQZWMnSN9/dl4N6dO5pjfM6OwkVHBtRBUyeCgkv3rjJ6vD4woHFD8Fihs5FDwJBBUyfDUjv7pHO0DwYF0f75rDihsZFDwNBBUiRfUczc0Id1b0V0+32AUvnD/gieCoEpCCgcudhju6fZ66gUGo/CF+xc8EQRVEjpHOnAW1mCY3Dirhh6fAoWNDAqeBoIqCV5sOLycYUzh61cpovAKcih4EgiqZHiOdCvmQNnAiqtO9kBhI4uCp4CgSkYUHjJY2FuRyxwWk0XBU0BQJYOgM7rTbgd53XOisJFHwRNAUCXDoDMM5jIHFF4CQZUsD7po5ioIChuFFHw7BC4mw9sAAA3cSURBVFWCwigsh6BKbgQNucpAGoWXQVAl00EHB4PdH6tBYSPDgmtIF9RpmvsO6jDjemHv7vRfmyc3Chv5FVxEsqBuC911UIdp54Znb9xef8aTorCBwpHxWuiegzoscvKmwnMnulDYQOHIVKjwnH61t/q2P2egsIHCsalvIL30+NDtXWEG0gtA4ejUN521+fjQ4O/NWQuFDRRWUkvQuAbPBIUNWrYSggpBYYOCKyGoEBQ2KPh2xoeROws6TmVBUXhfBZ+YzNlX0AkqC4rCuyr41CGVXQWdorKgKLyrgvdPH3Qf21XQKSoLisL7Knjw49vsx76CTlBZ0E0Kv/vLw8P33c0//4LC80gyndX2yJcfOws6TmVBtyj8+O33x7efPz/ffvul99C6TaLgawjs+3ofZX79WXzQuVQWdIvCb5+63fc/Wjf80zcoPJ+YQQPzV37vi8IJKFThE6ee+In3/3p+WfBwZsFfgE2YoOFFl+WDFWDvLKn3K9sBfvzu7w+fP+8Wr3tX4T17Of3DSL1v4LwuO9zn3GEqqiRKL/zqixfnn+/++vz4+M8XKDwP3UD6fC94bHj8aJMQKqokhsKv3J738bvuzrpNouBh5l/CfumAvc+l8B5pby/bgrVQUSURFL72wSi8kGVB5xvXBMbQvUmtw3DsLYSKKtmu8Gn0fOHtk8zv/sFx4ZksCjrXuKYZM9idn17wByNARZVsV/in8+Tz9+cDSq8eHtwued0mUfAQM40LGewtXPetm9ugoko4wdLIv+D93Vtnirl/LlbfTXfpsf/7CaCiSlDYKKDg/Ulmu+Mre9PgEoLGobKgKFxEwYO2etIGh9veKkUEjUFlQVG4iILfVPh6e6Qbttsbt3oFVFQJChtlFHwwkG58O1uDgxPS12F3co2pqBIUNgopeFjNgMH9fWZvx3k41tZCRZWgsFFkwb1hc6+r7V8w7Bts97Y+/1yoqBIUNooseGjf+HortJKncMK+mIoqQWGjzIIPB8yH0FVLoV3llONpKqoEhY1CCx441WP8KkNvCI3CAlDYoOCrCYyju4d6U1kMpOODwgYFX8vUCHmgMNNZ8UFhg4KvxJ2t6l3mcPAG0qmhokpQ2Ci54P7ZlUNTnYnqsY/yiLQlI1BRJShsFFzw4DGjwcVLg7UDf0AFFVWCwka5BW8NbYIKB++G/4AMKqoEhY1yCz6hcDu5NThXK/gHZFBRJShsFFxwbyDd/5CdsR3j4B9QQUWVoLBRcsHbE57bf3oGuysfRw44RdqWIFRUCQobhRd8pG9thlcWBoOKHaaiSlDYKLvgvRmrbtFQzFBQ9d4wFVWCwkbZBXcdtFuNv6hbFYWFoLBBwZfjG9zNavUePDCQloLCBgVfweDAb39i+rpiOCjTWVFAYYOCb+Lq6yKFtVBRJShs7Kjgg6sJbw+kxVBRJShs7LHgwcsK9xg0SGVBUbiygqeFoEpQ2Kis4GkhqBIUNvZS8JufLruXoDepLCgK76TgobOifXYS9DaVBUXhfRQ8fGWDxz6CzqCyoCi8j4KjcEdlQVF4JwUPGdy72DD+k96EiipBYWMvBR9OZ/WM3kvQm1QWFIV3W/B+r7zboH0qC4rCuy04CicEhQ0KHhUG0ulAYYOCx4XprGSgsEHBlRBUCAobFFwJQYWgsEHBlRBUCAobFFwJQYWgsEHBlRBUCAobFFwJQYWgsEHBlRBUCAobFFwJQYWgsEHBlRBUCAobFFwJQYWgsEHBlRBUCAobFFwJQYWgsEHBlRBUCAobFFwJQYWgsEHBlRBUCAobFFwJQYWgsEHBlRBUCAobFFwJQYWgsEHBlRBUCAobFFwJQYWgsEHBlRBUCAobFFwJQYWgsEHBlRBUyP4Ufl0OJW3rJgi6NyyoSOF17yq8ZyshqJD99cLrNomCKyGoEBQ2KLgSggpBYYOCKyGoEBQ2KLgSggpBYYOCKyHoLPyvg5wLChsFFnwdBFWyKWjvS5nngsJGeQVfCUGVbAnaNOscRmGjuIKvhaBKUBiF5RBUCQNpFJZDUCVMZ6GwHIIqSRbUsR2FjX0X3IGgSlIFdcfcKGzsuuAuBFWSKKg384XCxp4L7kFQJSiMwnIIqoSBNArLIaiSDUGXTUcznTWgtIKvhqBK1ge9dVB4/FEUNgor+HoIqmR10FunZk08isJGWQXfAEGVqBSeehiFjbIKvgGCKlENpFH4NoUVfD0EVSKbzmIgfZPSCr4agirRBWU66xY7K/g4BFVSWVAUrqzgadlb0PGB9LoLnLaBwgYtW0jBQW2/1zdzfD5r5WXG20Bhg5YtpNygZqtvZjM6J732wz62gcIGLVtIMUH7/jUOg4WhX0fhM8UUfCsEVbIi6EDAoMI3uuHlT7sRFDZo2UIKCTow8+pqcHnoLzCdZRRS8O0QVEk0hUMD5BGHOahkFFLw7RBUyUqFB0sOgZHz2FAahY1CCr4dgipZHDRk5WXBzRmty00UNsooeAQIqmRpUFfKgMqTDl9vorBRRMFjQFAlGxR2bfX64f7a/QdR2Cii4DEgqJJVA+n2Rtcf23xWfwLLO9aEwj3KKHgECKpkYdCzpdebl0NJ3nHhkUPG7Z3TTxQ2Sih4FAiqZFnQwZ5uT98xh9s75x8obBRQ8DgQVMmioMMJ5oDBg4NIgyUobORf8EgQVMk6hb1J5hsGD0FhI/+CR4KgSmYF7R8W6iaonONI4WF06M+hsJFvwSNDUCVzgvb3Zq+udtNZ13muwIT0YSgyChvZFjw2BFUyI+igdw2MmzunnfM+LncHfTMKG7kWPDoEVbJG4d7Vhb6s/smWwT1kFDZyLXh0CKpk4UDaWeIOp72htT9THXgHQGEj24LHhqBKlk1nOUsCXNb09D247wAcF/bIt+CRIaiStUGDBju9rtvz+jvMKHylqIJvgaBKVgftyerL65yH6f7GxWEUNsoq+AYIqmRL0PAe8ZjBKNynuIKvhaBK4ig80NcdR3vrn36isFFcwddCUCWbgoYNPoSOQ13WP/9AYaO8gq+EoEq2BR0x+MZHRaOwUWDB10FQJZuDdqNnfxJ6wmEUNsos+AoIqmT5Jf+BZc58lrMMhW9QQsGjQFAlGy7595c2Q4e9VZw7KGwUUPA4EFTJlkv+/aW+woeewe5dFDbyL3gkCKpk4yX/7tIRvw8D9VHYyL/gkSCoklUD6cEouZ2IDv8CCgcpoOBxIKiSNdNZg/H06MSV20F3S1HYKKHgUSCokjVBp4/6BtdkOitAMQXfCkGVrAo61+CRNVHYKKfgGyGoknVB5xocXhOFjYIKvg2CKqksKApXVvC0EFRJBIXf//jwxYvBTRS+CUGVVBZ0k8I/fXn6r38ThW9CUCWVBd2i8ON3z4/v/vaidxOFb0NQJZUF3aLwydmTu97NhzMz/wIACNiksLHuXYX3bCUEFbKnXhiFZ0BQJZUFReHKCp4WgiphOsuorOBpIagSDioZlRU8LQRVEkHhx2/P53P89E17E4VnQVAllQXdpPA46zaJgishqBAUNii4EoIKQWGDgishqBAUNii4EoIKQWGDgishqBAUNii4EoIKQWGDgishqBAUNii4EoIKQWGDgishqBAUNii4EoIKQWGDgishqBAUNii4EoIK2Z/CBVHNhwQRdG+MB0XhXULQvYHCBgXfGwRF4X1C0L2BwgYF3xsEReF9QtC9gcIAOwWFAYoGhQGKBoUBigaFAYqmDoXf/eXh4fvu5p9/ue/m6HDS9b7IfV+8PX+n5jfn23uu6PlbU5xKhopahcKP335/fPu5fQvU2y9vrFw0Trred27sj7eXxrzjij5+e/76hdEvUjlThcJvn96k3/9o3fDp6yj2S5eu/81Xu+P0vnxmvxV99fCfl68gHPk6szNVKHziUvH3/3p+a82CcdL1v39yd7y67jDst6L/88upihNfKnqmGoUvFX/87u8Pn++25k66vSt8HVXtu6Io3PHqsuP07q/Pj4//3Ov40km3d4Xb8eSuK4rCLa/c9+k9t+zjZLV3xCt3Vme3QVH4yitvJn63BT9zSbf36SxvEmu3FX3HdJZxGmtdOB2JePePnR5F9NLt+6BSuyu874qefeWg0lPw84kA35/fuV897PiMh2u6wBe574xLv7v3ip4VvlRyrKhVKAywX1AYoGhQGKBoUBigaFAYoGhQuHp++9PP994E2AAKV8GbD7527nzmPvSsaZpPUm8PxAOFq8BV+LdPXYWfffhff/rvP3w2/BUoBBSuglGFf/3DZ08D6WcfMZYuFhTePS+b5oP/OCl8GjJ/9iRt0zwZa3cuCp/W+vCHJ7kZUZcHCu+dZ09uvmmeFH728ZOoTz/PvXB356N/nxX+/atPji/pjAsEhXfOr+f93GcffP3bH7+2eyeF2zvnTvnj04pvPvyvP359449BhqDwznnz1Am3+8JvToPn677w+c7xtG9sU9LPTGUoDBTeOS9bhZ/2iT/8v5deuL1zPB8XPg22n5RmXrpEUHjntL3wZdhsCrd3jmeFf/1fXz/tDP+f86pQGCi8c8zTlx98/eba054Ubu88uf2k8LmT/uj/fcWEdIGg8N45zTufZqStA37a6z0dOnLufPTvP/370487saEwUHj3XI8Ln35+/ex0dKn56Of2zu9fNecp6dPZHb9/xYRWeaBw9XCZQ9mgMEDRoDBA0aAwQNGgMEDRoDBA0aAwQNGgMEDRoDBA0aAwQNGgMEDRoDBA0aAwQNGgMEDR/H9U8tjUWPyNXgAAAABJRU5ErkJggg==" alt="" />
笔者用k-means聚类效果如下
# 用K均值聚类
model1 <- kmeans(data,centers=,nstart=)
library(ggplot2)
p <- ggplot(data,aes(x,y))
p + geom_point(size=2.5,aes(colour=factor(model1$cluster)))+theme(legend.position='top')
aaarticlea/png;base64,iVBORw0KGgoAAAANSUhEUgAAA8IAAAKfCAMAAACSUoECAAAA/1BMVEUAAAAAADoAAGYAOjoAOpAAZpAAZrYAsPYAv30Av8Q5tgA6AAA6ADo6AGY6OgA6OpA6ZmY6ZrY6kJA6kLY6kNtmAABmADpmAGZmOpBmZmZmtrZmtv9/f39/f5V/f6t/lcF/q9aQOgCQOjqQOmaQkDqQ2/+Vf3+Vf5WVf6uVkP+VlX+VlcGVq6uVweujpQCrf3+rf5Wrf6ur1v+2ZgC2kDq2tma225C2/7a2/9u2///BlX/BlZXB6//Wq3/W///YkADbkDrbtrbb/9vb///l5eXna/PrwZXr1tbr///y8vL4dm3/Yrz/tmb/1qv/25D/68H//7b//9b//9v//+v///8Xo35LAAAgAElEQVR4nO2dC3vcxpWmGY1HFOm9r5TsRrRj7Y6TMBnLjrU7u3Z023FTY0nUhVL//9+yRB90o26410FXod730SN2A+gGvj54WReA5MkWADLm5NgHAABzQGGArEFhgKxBYYCsQWGArEFhgKxBYYCsQWGArEFhgKxBYYCsQWGArEFhgKxBYYCsQWGArEFhgKxBYYCsQWGArEFhgKxBYYCsQWGArEFhgKxBYYCsQWGArEFhgKxBYYCsQWGArEFhgKxBYYCsQWGArEFhgKxBYYCsQWGArEFhgKxBYYCsQWGArEFhgKxBYYCsQWGArEHhlHgdRGWbBXeV2iGvDBROiQx9yPCQVwYKp0SGPmR4yCsDhVMiQx8yPOSVgcIpkaEPGR7yykDhlMjQhwwPeWWgcHw+/u6nia/M0IcMD3lllKzwy5OT+33bfHzwsG3Vp0d3fnIeP6n+vz45OXkY2qjtMJotmpPuzZs3zgn4/t//ZG/z4sUL+yR9/+V+x4fFjx8/tre5PbjffGvv6vz8vG9Xr+/du+ft6otf7G3Ozs6sbT492n++h8Wnp6fW4VQfVb3RYfHl5aUXyz3kq6srdxM5GmObt2/fhmLdHlVvRfKiYIU/PqjLbvD+PzsLXsrZE8JTWM6O91/+l9/90+FVAYXdfbz/8v7+4eGce/PGcLg+2ju2Vy9emA5v5bvNtX2yP35sOlzt6999u73evdFh6fm54bAcxhNnV5XBjcPVmuu73iGfnRkO797li1+q3RnbnJ6aDtdvIIdz2Oby0nS4jvXSPuSrK8Ph3Sb3ty/v2ofz9q3hsPEJPrlb/VsTKGzy/kunup8e+Zo362yFPz44qRS+Pvm90ZH2Fe7ax/6Ue/PGdLha8/LkP9lN44sXlsO3K65v3+bTo4fGNo8fWw7vc//2W2NX5+emw7sNrv/BaYXv3TMdrtY8ub91DvnszHT4sBtzm9NTy+H6aB5YHYfLS8vhrbSd9iFfXZkOW5sctnn71nS4+QSrzfadjJVQrsJVP+8331Y9wqrlqr7erRbd+Un+vz23/uHLO/9y69v7L//xdsnvH+x6qfXKavt/fLTbatej3Cn8u//7aKewGLB7x9rtL36pvlav/eJfd29Qv2y3j5+eHNrsdoX/3y/v+xSuqIXoVPjl7jtGu8Kf/ue/9Cn86X+4evoKG6J0KfzS7gH7Clffla6tQ56msHyCxqZroVyFd63wxwd3D1+vT+5XLWTV3xLxbs26XXbr4p2f3n/5m29l4X7lbS/xS/m6N7TpTlffFOp3NBV+efLw9p/sQ16228dt+7AfOrcr7A1Qwwq/tAaoIYWrJOaufIVf3nfHwp7CH3/7H+rxabvC11/8L2doHlK47jZ0KFx9oPaw21N419c+Mb95BRSWTxCF14R0pKVZvZYToNKrsrb6b7e6OjH2zlUW7ldWXw/PX4rezaRW5XD9jrbCd8193L5MDuH6MKu2PwEDY2FvjskbC2+bmbH9Um8sXNEzFv7439wG3x8LV2Pcet79sI07Fq5SOUPzwFj40FbvlwbGwvf7xsJVj+e/WkOI0FgYhVeHNIOVq7sGslpU6bV72KbwfmX1VZ7veGgp/PF3/7Rrb6t3dDvS0grvX9aq8KQZ6Wbu7bDYm5HeumNhb0b6ycMBM9KHt2mfkfaH5v6M9HY/C9UxI+2Phf0Z6dDhhGekUXhNVP5UbeX+63ZqK7zdy3tQ+H+fPAy0wtvqpXcP+9juOwIhhU1kne+Vu01zdarrfTwf7E2qUfr+qtiAt2nfpjpeV2H/fQ7zYp3v033I5sfTtc17prNWhqhbNYli2U5bcyx8q5eMhRuFzZXXJ2YDe1D45cnvq4tK9TvuV1YD5ydVE3x/P/w+7CMwFm47ATu3qa/gdG5TdaLf/8df5u6qeZuO97m107lg5G9zGAq3b/PpkdeRdjepvHxyt/tt9rG4qLQidv7cjlvvVFdxZf749mk1rq1npG9Xv9/NSDfOBWekd4ubVviJNGLNjHR1B8M/7mekf9nto36ZKOzPSLeegF3bPBnSfL48OenTalCDv3+bTj29O028bZo+bfs2gVs73E2uT/xbO1pi3X7y3NpREl3XhVsZeYNl4Lpw6ATcxt1mwV2ldsgrA4W7ean/Lbuacq3J0IcMD3lloHA3HfdIx+I6dI90Nj5keMgrA4VTIkMfMjzklYHCKZGhDxke8spAYYCsQWGArEFhgKxBYYCsQWGArEFhgKxBYYCsQWGArEFhgKxBYYCsQWGArEFhgKxBYYCsQWGArEFhgKxBYYCsQWGArEFhgKxBYYCsiaDwZhKvp71sHhOPdRYE1aSwoChcWMGXhaCaoLBQWMGXhaCaoLBQWMGXhaCaoLBQWMGXhaCaoLBQWMGXhaCaoLBQWMGXhaCaoLBQWMGXhaCaoLBQWMGXhaCaoLBQWMGXhaCaoLBQWMGXhaCaoLBQWMGXhaCaoLBQWMGXhaCaoLBQWMGXhaCaoLBQWMGXhaCaoLBQWMGXhaCaoLBQWMGXhaCaoLBQWMGXhaCaoLBQWMGXhaCaoLBQWMGXhaCaoLBQWMGXhaCaoLBQWMGHcX5+HmWfyQeNBQoLFFyTEUHPzyM5nHrQaKCwQME1cYJ2SHp+HsvhFIIuAgoLFFwTO2iXpCg8mgwV/vDn59vt558vvn4uz42HKNzL8YN2WorCo8lP4ZvvK2GffVP922E8ROFejh+021LGwmPJTuFXF/982wrf/PBUWuOt+RCF+zl+UEPhkK3MSI8kO4X/7dfK1w+1xtut+RCF+0kgqGlwJF8DJBB0GbJTWJwNKnyxY8A7wBGpG+Hzw6Nm+VGPC+IwS2Fh2ncVvmdrYgQ9P9D0qA9PYu7z2EEXY02tMAoP4MhBz33MZRH3SUU1ma8w01lTSUjhjWdwTImpqCYRrgtzUWkiKSm88Q2O5zAV1SSCwjff7+7nePbd4SEKD+LYQV1XAx3rOBw76GJkqHAP0w6JgkfFMXHrruowGIVHgsICBY+Jq2J70E6FZ/lMRTVBYWGtBfdcbA3aZvC+qZ7hMBXVBIWFtRZ8iMKyvt3g2f1qKqoJCgtrLfgAhesNfIPP3UnqqQdBRTVBYWG1Be8dCwc60K1MPAYqqgkKC+stuGOeJ+IYhSc6TEU1QWGhlIL7Ig4zd5bDVFQTFBYKKbgp4rnxY4b9be8ch6moJigsFFJw18nNaIXPm3cavFcqqgkKC4UUvNFwgLmNtefm3HTzRkP3SkU1QWGhlIL7reoYj91XD9wpFdUEhYViCm41o66dgxzejB4YU1FNUFgorOAby2H7Sbu6biOMwg4oLFBwTUI/L+w8aW2G3SZ86D6pqCYoLBRW8BrT4EbMdoXtFw7e5/GDLgQKC8bncO/evZYnsSms4Hu8RrbLYeeFw/eZQNBlQGGh+Rzu3TO0tZ5Ep7CCGwx2ePo+0wi6ACgsHD6He/cMba0n8Sms4A2hZjiywWkEXQIUFlBYk2kKz9tnEkGXAIUFFNakXWH7WUSD0wi6BCgsDBgLx3e5sIIb2JpGlldII+gCoLDQOSO9/y+2w4UV3MQVNbrBqQTVB4WFls/hnkvUxriwgnfSDIw3cf5AaapBo4PCFYah4eXNBjEb45IK3mulsUGUthiFNUlNYauVbVnhNsYxKKjgY6xkOmsUKOyI2r6maYSt6a7Zn8OyHKfgo6xE4VGgcJfCG3P5vcOkljlJPcfhcgqOwnqgcL/C+7bWGwvP7VWXU/BxVjIWHgMKb/rHwuaTQ9/Z7l3P+ByWJYOxMDPSo0DhirZpKnOps4k3Sp78OSxLsjPSsUFhTZJTuO1zCCncdKpnO1xYwZeFoJrkonCbsnbXG4V7MYMu1RwfPehSFKvwPfvKUOhz6Gx0vXWTP4dlOfaZHfEWym6OHXQxSlXYsG730P0c+rrNfg978uewLGn87qwF9onCmhxf4b5G1DG1TWFa4WGgsCIoHDKwXdl6W8Nhq0We8jksCwprUlhFc1TYus3Da4+nfQ7Lcuwzm7FwbApVuGcs3K7wvZYNZnwOy3L0M5sZ6ciUqnDPjLTbX753z754FMXg0gq+LATVJAWFHQIz0puW1ngTWDHnc1gWzmxNCguatMI1vQ7P+mGlwgq+LATVJEGF2zwM9KTNrnTHKwdQWMGXhaCapKdwsC11hsOBH1ya/hHsKKzgy0JQTZJTODiitZtepxUODoHHSl1YwZeFoJpkobA3Am4bFtuvGP85LAtntiaFBc1L4faZra43GfI5LEsCZ/bZ2dkC+0wg6DKg8I62frShKgpPxwp6draIw+Ggl5eXmjstrKIpKRwYxrqmDutHo3AIM+jZ2TIOB4NeXnY4HEHvwiqalMLe52B2oMXwHoPDM1wDPodlKVrhy8sOhzv1HkhhFc1CYeOJNzqu11gvmPA5LEvJCl92Kdyp91AKq2g+ChsNstU2m2Po8d3o4gpes4zBARsvUTgaGSkcan4PS3p/c8Cgz2FZFim46+jyM9KWjvKw22AUHkUGCvsD3w6MF0z4HJZliYJ77eyyQS8dWeuH3QYzFh5F+gqPMtjoSU/5HJYlWsHb21J/tLto0EtH4f1jV2uvs82M9HBUFX4dg1EGN68ZuZc4x3ocak+HrVoyqGHwpbmgfuBu9PqwLAY5V3QUElRJ4WnfVQa0wsHbpUe3vP63smWJ9D377Mxva511xpIlg7qNsNH6uot6RscTyLii40i/I33oFlu95Xv7S8STLiH5ZFzws16HzQXHUdha0rJNZIczrug4MlDY/Otn4R9Bmm9wzgW3FXZldp8sPxa2gvYaHM3ijCs6jhwUXoKcC24q3NEgL3ZLtIH42B40aHAch3Ou6CjSVzhCEzuArAvuG9w1MF7IYsPE0QrHcDjrio4heYWNga6mzLkX3OpGN5q63ejlJDZNbAnqeYvCU0hdYeNyb4xZq1ZWUnBHUlvXswUdtlRs/WFDv+VF4dHko/C9KDPPbWRZ8ICKvsHHcbhfYWWD86zoFFBYyLHgQRWdbnQ2CvuvmkmOFZ1E6gpv7H50sQq3N7jdL2qmuWyDl+lJ149fh7R0DW4dCE90OvWKRiN5hZtJrJLHwgHrQi46Cw6bnHnMOe4hWC4Gm9ZWgy873mkEiVc0Hukr3FDujHSbrc5CZ5kv7tm+NY5x+F1YMgbM9DrRXp86+E5jSLuiEclJYU3SLniHwodn3lA3bPBiF4W7FPbHwSg8GRQW0i54u8LGE3esm67CjsGuwqH2evwhpF3RiKCwkHjBA/L19ZjbFF7Q4eCTjeGrscZ4HnB4whEkXtF4oLCQesED6jWLgtIetxm25pFfh1vW2lnzBaGGeJrByVc0GigsZF3wcLu7l/w4CpuYQf0+tLEm5i2WWVd0DCgsZF3w7uZ2//h4EhtBA8Ngc43v8GSbs67oGFBYyLngbYNe/y7LIzncBHUNvuxbM71Fzrmio0BhIeOCtxpsda3tjePseSh1UM/SwCy1Mxqe0avOuKLjQGEh34IHdA23yGfGFFeUPQ9GggYMtjvM8sxvhFG4ExQW8i34CIObraLseTC7oC0GH34DrS3u5vAIhXtAYSHfgg9U2PoaZcfD6VbYEnnjTGBNNjjjio4EhYWMC262sr1sthoG92g2ROGDqo60Uw3OuaLjQGEh44Lv29WBCsfZqUVfUxkaC/tLLs07LecfVMYVHQcKC/kWvBndDmiMVYL2WheYkfaXNN3pOBbnW9GRoLCQb8Ebha3n9VLX4Nug0XvSQxVu+9XwLQrPdDjfio4EhYV8C36w88y5f8M3WOmi0mCF/b9o2N6Tnu1wvhUdCQoLGRfcaWaNy0cbR12teyz7fLPuzvJe6DqLwuNAYSHngoccttbsl7nbRaNHt9ctG1n2BhrnWceUc0VHgcJC3gVvU9Ne6Km+FM1FJXu53wJba+btNO+KjgCFhbwL3qqmtexYBu+ChlxtVzjC3wjPu6IjQGEh74K3q3lY7M5rLYmtcHMPVofC88m7oiNAYSHzgneY6U5kLW+wq7Az76xjcO4VHQ4KC7kXvMtfC5UbLHtwb7BcwuDsKzoYFBbWWXDX380xg7YprLTTdVY0AAoLqyy41wRvjh3U6UdHmLVqZZUVDYHCwioLbrh7GAIfOeilef+G6k5XWdEQKCyssuBm67sfAqcSVNvgZIKqg8LCOgt+aHubOax1Bg1QWFAUXmnBG4P3Dq80qE9hQVF41QU3rwavOmjN48ePywi6A4WFVRe8MIUfV5QQVEBhYdUFL0vhx0JX0F0zrQAKCygcnaLGwn0K1+s1do3CAgrHp6QZ6VaFH9soNMYoLKCwJgUEbRkLPw4Sdc8oLHBma1JC0OCMdNjguA6jsMCZrUl5QZteMwqjcFwIqsk+6N7U0QpP0BuFBc5sTUoL2qhqajtA4SlNNAoLnNmalBPUa21NZXsdntTNRmGBM1uTYoJ6HWbjKwqjcBQIqkHtnT/kbXvW+jYojMI9EFSBvXiewRvnebfBjIVReAAEjU+wza1VDC/oequxe0dhgTNbk1UF9S1raXQPyhqNb18zPIFMFX53UfHd7vGHP15c/OFXFB4GQWfiG2ho6cxIe5v4a+aTqcI7jb9+Ll+/sRZPOyTObE1WFNRX0LTSuy682XiT0Si85+b7H+XBs+9QeDgEnYenoCXlduP2nMO3WsY8onwVflV3nj///SkKD4eg82hTWJ5s98oeXHW+Bl4994iyVfjzz3UjfPPDXy6+2lm8Gx9fDP8mADCenYPu82bJ/qnTcd4aC7bN5u6brYDhCn/48/P6wZ+ebm/+9vywYtp3FRonTVYV1G04zXZ123GBKdCVjtKtzrYVfmVOYt380HSmpx0SZ7Ym6w5qSGgo7HWdnXZ5syldYWsSC4UHQ1BNtkaj7BjrGIzCh6Hw7trSh79yXXggBNVkPyO9w5XWNjnabR65Kly3u1Vb/Ori4utmKIzC3RBUicfOL94JCWwpHMngbBVuZ9ohcWZrstKgpn9i49Z+3tGPjiUwCu/hzNZknUFNBWsh7d+dZYmKwigcB4LGwnKwU2FvgdEszz8OFBY4szVZZdAWhf0219zAbZYjHAcKC5zZmqwyqN2MyuNt4Kf/ra3dhjnCcaCwwJmtyTqD2s3o7vG2MXTfW7a29vvW8w8DhQXObE1WGvSx9zeSHIUfO1v7BkdwGIUFzmxNVhzU1rBR2Otme7/AEoVReBIEjYrjoTcWNrZyfhMACqPwNAgaFU9h5/fRbmxxPa9RGIXHQtCo+Apbazbu1LTpLNNZKDwFgsbFGQvXiw5f3Dks97URjgCFBc5sTdYc1PKw/kkl977LSO1tGBQWOLM1KSloYMZK1WAUruHM1qSkoEGFNfeJwgJntiYlBQ0prLpPFBY4szUpKqg3FlbeJwoLnNmalBXUuXCkvU8UFjizNSGoIigsUHBNCKoICgsUXBOC7lHoWKOwkGbBFSCoJn1BNaa3UFhIsuAaEFSTnqAq15hQWEix4CoQVBMURmF1CKoJCqOwOgTVhLEwCqtDUE2YkUZhdQiqSWFBUbiwgi8LQTVBYaGwgi8LQTVBYaGwgi8LQTVBYaGwgi8LQTVBYaGwgi8LQTVBYaGwgi8LQTVBYaGwgi8LQTVBYaGwgi8LQTVBYaGwgi8LQTVBYaGwgi8LQTVBYaGwgi8LQTVBYaGwgi8LQTVBYaGwgi8LQTVBYaGwgi8LQTVBYaGwgi8LQTVBYaGwgi8LQTVBYaGwgi8LQTVBYaGwgi8LQTVBYaGwgi8LQTVBYaGwgi8LQTVBYaGwgi8LQTVBYaGwgi8LQTVBYaGwgi8LQTVBYaGwgi8LQTVBYaGwgi8LQTVBYaGwgi8LQTVBYaGwgi8LQTVRVfh1PuR0rLMg6NqQoEoKT/uuwvdsTQiqyPpa4WmHRME1IagiKCxQcE0IqggKCxRcE4IqgsICBdeEoIqgsEDBNSGoIigsUHBNCKoICgsUXBOCKoLCAgXXhKCKoLBAwTUhqCIoLFBwTQiqCAoLFFwTgiqCwgIF14SgiqCwQME1IagiKCxQcE0IqggKCxRcE4IqgsICBdeEoIqgsEDBNSGoIigsUHBNCKoICgsUXBOCKoLCAgXXhKCKoLBAwTUhqCIoLFBwTQiqCAoLFFwTgiqCwgIF14SgiqCwQME1IagiKCxQcE0IqggKCxRcE4IqgsICBdeEoIqgsEDBNSGoIigsUHBNCKoICgsUXBOCKoLCAgXXhKCKoLBAwTUhqCIoLFBwTQiqCAoLFFwTgiqCwgIF14SgiqCwQME1IagiKCxQcE0IqggKCxRcE4IqgsICBdeEoIqgsEDBNSGoIigsUHBNCKoICgsUXBOCKoLCAgXXhKCKoLBAwTUhqCIoLFBwTQiqCAoLFFwTgiqCwgIF14SgiqCwQME1IagiKCxQcE0IqggKCxRcE4Iqkr7CHx+c3EXhaBBUk8KCDlR4u315cnJyH4XjQFBNCgs6WOFxFk87JAquCUEVyURhkfjOTyg8F4JqUljQEQpf3/r7cPvp0Re/oPBMCKpJYUGHKvzxwcmJuHvdNMMf/nhx8Ydfdw8//3zx9XMUHghBNSks6ECFPz74zbf+0nffHB4++6b6h8LDIKgmhQUd2goHefbd/tHND0+3H/7cNMPTDomCa0JQRTJV+PPfn+4fVvpWGqPwIAiqSWFB5yh888NfLr566ih8sWPgOwCAAkMV/vCnp9ubv+16z7TCoyCoJoUFnaNwRe0tCo+CoJoUFjSSwkxnjYKgmhQWdI7C775+vv3wV7kuzEWlMRBUk8KCzlF4++pidz9HdW3p5ntu7RgOQTUpLOgshduZdkgUXBOCKoLCAgXXhKCKoLBAwTUhqCIoLFBwTQiqCAoLFFwTgiqCwgIF14SgiqCwQME1IagiKCxQcE0IqggKCxRcE4IqgsICBdeEoIqgsEDBNSGoIigsUHBNCKoICgsUXBOCKoLCAgXXhKCKoLBAwTUhqCIoLFBwTQiqCAoLFFwTgiqCwgIF14SgiqCwQME1IagiKCxQcE0IqggKCxRcE4IqMiLoixcvBiwaAgoLiRc8HgRVZHjQFy88YQOLBoHCQtoFjwhBFekOahj64oUnbGDRMFBYSK/gSmQS9O3btzN3Ov8dxtMZ1DQUheOTyZk9nzyCvn071EB/O1ky/B0i0hXUVPQFCscnjzM7AlkEfft2qIH+drJk+DvEZKDCIYM3Uw1G4ZoszuwYZBH07VAD/e3eWozd8UyGKRzXYBSuyeLMjkEWQdeo8MZthP2Vt8smaIzCQhZndgzyCDpb4fTGws2MdIfCU5piFBbyOLMjkEnQuWPhBGekG9ons6Y4jMJCJmf2fHIJOndGOuWgB1P9K8UoPJmECx6X3IMelO1zPOGgxsDXcnjahSUUFhIueFwyD3roOPf2tJMI6vtomNrWnx65UxQWkij4EuQd9DB91T/flULQlmmrjpEvM9KTSaHgi5BtUNtc76qR53MCQX1HrS70tAloDxQWEij4MqQfNNC4vnUwlphbWK9JIKjn6IuAwvsVwTccYjgKCwkUfBmSDxroIPsGuw47bXL18HhBD965M88v2hRua40HtdIoLCR/Zsci9aCBQW7IYLsrHWqSjxbUFrM2+PDFeWg+89rcYT1tFBZSP7OjkXrQfoUDWwbljnDgY6mCNt45Ta51b7SpsOlzYOCMwsNI/cyORupBB7TClqr2Fm1vsRCdCpuWuotCDoesDoHCQupndjSSDxrQr8NhdwvjyezDHo+tsNuBDs5mtSo81GAUrkn+zI5F+kED9nmt7Vt3W9/h44+F/WmsIezfa6jBKFyT/pkdicyCGvJufIUdm82udDIz0ptxCh/ea6jBKFyT2Zk9nbyC7q2Ur25PunMEfPygjYSOpcMsHmgwCtccv+ALkWbQlqHrXsvQYNgbGjsD4OMHNYx0FG2V13V4yE5RWDh+wRciyaBt00/tCgecdr4PHD9om8ItDm9C498BGqOwcPyCL0SKQd0+sLdigMLeqxMI2mJwe//ZU9gROggKCwkUfBlSDNqqcPtYuN/hFIKOMdi6/NS8vt9hFBZSKPgipBi0ReHa2sPj0AXi9q50QkH9Zra7GQ68sAMUFhIquC5JBm0zuP9nliyFrVckFNQw0WluXYU3trEoPIKECq5LmkHbZXVX9xl82DKloI24G1tV3+HgCztBYSGlgquSTdCDmm43u6sbnajCm6bpdRd3K8yM9HCSKrgmOQQ121RL0sNq2+7gTyglcHeWTcDT/aJBHeY2UFjI4cyOQgZB7X5xh8L7G6I3pvX7rVK4R9ri4Gmz2lw0+TfwoLCQwZkdh/SDuk2s53DbojRmpLvHtc6vnjWnsabuFIWF9M/sSKQftFVh94eS/C500go7v73SWDQPVYVf50NOxzqL9IPWPr5+62GtN5YFNjpe0NrQllWH9bstOrYdjgRVUnjadxUaJ00yCLpvXpvecXsrHKB+lzTGwn4ja05Az5rG2kNHWsjgzI5DDkH3HjY+mm723W1Zb5XEjHTIUBTWIYczOwqZBjVHuS13WyZzUakhrKjr8MydorCQQsEXYUVBvcY3E4XNH2eYbzAK16RQ8EVIP6jV4PZv6ylcvzqFoL0XmGKAwkIKBV+E5INaw95hW+9fEV3hN2/ejNo+9JcNu2/ziAAKC8mf2bFIPag9oB20vfsLLmVFhKBv3ox02A/a5mk8g1G4JvUzOxqpB/VmlscQdyz85s1Yh0cEjWYwCtekfmZHI/Wg/tWhzYjhsbHhkKDdgqoqHA8UFlI/s6ORfNCAw5Pa5QFBewxFYRTugaBBPIWn9a37g/YqGmEsvAAoLKR/Zkcii6BtCo9xeKTCQVlnz0gvAQoLWZzZMcgjqPhqXytSVXj/yJN2lMVjgjIjHZk8zuwIZBL07f6GyfqJhsIbx+Dbx17XeVxfetyMNNeFo5LJmT2fbIIG7tmIPBY2mtg3FhpMn5EAAA45SURBVNYGYxweHjTivR0oLGRzZs8ll6ChaelRbzA06M5QFEbhKRC0nVl3eOzUGxi0VtToRzd9a0PtYW+GwgJntiaZBJ1t8EDrDoZaBr/xHg17N8bCAme2JpkEnW/wMOuseSz/yUiHmZEWOLM1ySXodIG7FTaWu+qG6Hk3B64LC5zZmmQTdLrBXdIZK4wWt8NkFJ4AZ7YmRQTtNlhWGb1md0Y65PCg/aKwwJmtSRlB3/gz0lZju3Emsjo71MPv0EJhgTNbk1KDuk2re0NWawd6sL8bFN6TQsEXgaCa2EE9Qd1LwW0bjHIYhYUECr4MBNWkU+E+3BcN3CcKCwkUfBkIqsl0hQMvG7hPFBYSKPgyEFSTzrHwUIVH3iSNwkIKBV8EgmoSmpH2RQ4ucl7GWHgsSRR8CQiqSWtQR9eeZnjUz/yjsJBWwRUhqCZdQcMKtw59B1uMwkJyBdeCoJqEg76xrxcZj83FXjvM3VmjSKjguhBUE386a2P9iixXW/Pxm+YV3CM9gRQKvggE1SRwUanz/g33SaiV7gOFhQQKvgwE1cQKGpizCjTCgZ/1H3dpGIWF4xd8IQiqyTiF2zYcd3MHCgvHL/hCEFSTaQp7Pw8xxmAUrjl+wReCoJr032BprmjZcuTtWSgsJFDwZSCoJr03WJorWjYMre8ChYUUCr4IBNUkcIOl0Tm2V1gPHcXH3J6FwkISBV8CgmriBe1vTq2+85R9orCQRsEXgKCa+EEHGRz+42gDQWEhkYLrQ1BNeoIGLB01cxUEhYUUC64CQTXpDupPQ6MwCo9muaCnp6f7h+sOatAZ1J+x6vlts8PURmEhvYIrsVjQ09PG4VUHNeh2LnBd6fD7aFu3798pCgsoHJnTU8PhNQc16HGuReFhm7eDwgIKR6ZAhYdJOW5rFB4MCkcGhcNbOJvPejsBhQUUjk15Y+Gxk8sDb/roA4UFFI5OeTPScy8PBd5vwFYoLKCwJqUEjWvwQFBY4Myej9HsOqwsaDuFBUXhdRXcHPw6rCtoB4UFReFVFdyagnZYVdAuCguKwqsquK2wLfOqgnZRWFAUXlXBLYUPD+XLqoJ2UVjQWQp/+OPFxY/Nwz/8isLDiBrUamtdg6sn9Zfsgw6lsKBzFL75/sftu6+e7h6/+8ZaNe2QKPgEnMHvXtpG4f3X3IMOprCgcxR+d9vsfv5ZmuFn36HwcCIGDU1g2eqi8AJkqnBF1RLf8vnvT+sFFztGvAPMohY0tKheEdgCVs4YhV/JAPjmh79cfPW0WTztuwrfs8fjtsKnBw6jZPtZvSTeEXRBRTWJ0gq/+vr57uuHPz3d3vztOQoPY2TQTuPaDG4mtYyhceglmlBRTWIo/MpseW9+aJ5MOyQKHqTHuD6DjWlpo1VexmEqqkkEhfdtMAqPZFTQocadWjjLUFiXTBWues81725l/vBXrgsPRENh39+NqzUKq5Gpws92k88/7i4ovbq4MJvkaYdEwUP4YtZ9Z2tp0GBzcsttm+clGAgV1YQbLIX0C+6paXWN3S5z2GBmpHVBYYGChzlttdXtHDvNq7VwMW8NqKgmKCxkUfAehX2l3S03G3/tAlBRTVBYyKLg3QqHmmVny01g7QJQUU1QWMij4JZ8PQa3bGe9IFKQPqioJigsZFLwlg6y1Y32f+K/nSg5eqGimqCwkGXBLRVdmf2NjuYwFdUEhYU8C+67unFGzN4PDKOwIigsUPDhnHod65afWnINXvLmLCqqCgoLqyh4S/t6enrqDpw3TguuCxXVBIWFNRS8tYvstsLNUveRDlRUExQWci641a56DgfGwv7rdR2mopqgsJBxwQ8GBhX2Rr+er/qjYiqqCQoL+Rb8YKBjsNsyO1sH3kANKqoJCgv5FrxD4fZGOfgGalBRTVBYyLfgrrnBnrM3MvbfIcrBhKGimqCwkHHBWx0OGByeu9J2mIpqgsJCzgXv6D37cm4Dsmp3pamoJigs5F1w00G3E72xr/uGgqJwJFBYoOCjMRwMtMCWnyisCAoLFHw8TiMc6lnXWwaDMhaOAwoLFHwCp+5Vo1EKK99iSUU1QWFhPQW3us/+Dx+uJ2gPhQVF4RUV3GyQ91+7x8LaUFFNUFhYdcH7ZqS1oaKaoLCwloKf9v1eu7UE7aWwoCi8koIH7+awWEnQfgoLisLrKLh9PTjIOoIOoLCgKLyOggcVtn1eR9ABFBYUhddR8JDCjtDrCDqAwoKi8EoK3mIwM9LLgMICBZ+BNyONwguCwgIFjwkKLwgKCxQ8KoyFlwOFBQoeF2akFwOFBQquCUEVQWGBgmtCUEVQWKDgmhBUERQWKLgmBFUEhQUKrglBFUFhgYJrQlBFUFig4JoQVBEUFii4JgRVBIUFCq4JQRVBYYGCa0JQRVBYoOCaEFQRFBYouCYEVQSFBQquCUEVQWGBgmtCUEVQWKDgmhBUERQWKLgmBFUEhQUKrglBFUFhgYJrQlBFUFig4JoQVBEUFii4JgRVZH0Kv86HnI51FgRdGxJUSeFp31X4nq0JQRVZXys87ZAouCYEVQSFBQquCUEVQWGBgmtCUEVQWKDgmhB0EFdXVxNehcJChgWfBkE1mRX06mqSwygs5FfwiRBUkzlBr66mOYzCQnYFnwpBNUFhFFaHoJqgMAqrQ1BNGAujsDoE1WSxGWljUxQWMiz4NAiqyVJBzQYbhYVVF9yEoJosFNQaNqOwsOaCWxBUExRGYXUIqsmMoGOmslA4QG4FnwxBNZketG862l7LWNgns4JPh6CaTA7ad1HYXcuMtEdeBZ8BQTXRUrhrNQoLeRV8BgTVBIVRWB2CaqI1FkbhfjIr+HQIqonajHSH4Sgs5FbwyRBUk0hBA7p2KDztNwXMA4UFzmxFMg4qTW7gMlJQ1Yk/HTEPFBY4sxXJN+hVg7esdesYOx4BCguc2YpkE9T17+oq4LAvtbNmwo7ngMICZ7YiuQT1BEThaYeUS8FnQ1BNxgf1DLwKKtzXk55ysHNAYYEzW5FMgrYp7Dnb2g4zIy1kUvD5EFST0UF9L12Fex3murCQR8EjQFBNxgY1rdzLWS/wOtOhBdUXFBayKHgMCKrJRIUPDzfGo3aFrc1QuCaLgseAoJpMV9j088rG2dh+AQrvyaLgMSCoJiODuo1wEG9zFA6SQ8GjQFBNxgX1usU9Clt9axR2yKDgcSCoJqOCmh3nDoeDL9kwFnZJv+CRIKgmg4I6Da8/Bd3RkTbvx7xiRtok3YJHhqCaDAl68NIxdYjBoZs4UFhItuCxIagmA4IaZnaqaztsq269IQoLqRY8OgTVZJzCoUHwxntkvCjQNKPwnlQLHh2CajJSYWvJKIyXo7CQasGjQ1BNRo2F7SUBTXsUvmI6yyTZgseGoJqMmpG2lgxtfRuF919RWEi34JEhqCZTgw7Tt9l0w60dHlkVfA4E1WTur4LvN/jQiKOwQ14FnwFBNZkTdLDD1uYovCe7gk+FoJosqDBjYYfsCj4VgmoyK2iruGGFmZG2ya/gEyGoJuN/2NB5Gm56WxwWUFjIoeBRIKgmE37YMLCoV2HrZSgsZFDwOBBUkxk/bOgsDSt8ZV0RFlBYSL/gkSCoJpMU9lpYfw6rMdi+d7oChYX0Cx4JgmoyRWHHSEts+2Jw+AUoLKRf8EgQVJMpY+HAQNf+rbT2GhRuIYOCx4GgmkyZkXYVDtxD3awI9LxRWMih4FEgqCaTgjoGD9jyihlpn3wKPhOCajIt6FCDw1uisJBRwedBUE0KCzpL4c8/X3z93HuIwr0QVJPCgs5S+Nk31T/3IQr3QlBNCgs6R+GbH55uP/z5ufMQhfshqCaFBZ2jcOVs5a718GLHwHcAAAVmKSxM+67C92xNCKrImlphFB4AQTUpLCgKF1bwZSGoJkxnCYUVfFkIqgkXlYTCCr4sBNUkgsI33+/u53j23eEhCg+CoJoUFnSWwu1MOyQKrglBFUFhgYJrQlBFUFig4JoQVBEUFii4JgRVBIUFCq4JQRVBYYGCa0JQRVBYoOCaEFQRFBYouCYEVQSFBQquCUEVQWGBgmtCUEXWp3BGFPPrCQi6NtqDovAqIejaQGGBgq8NgqLwOiHo2kBhgYKvDYKi8Doh6NpAYYCVgsIAWYPCAFmDwgBZg8IAWVOGwh/+eHHxY/PwD78e93D0MNI5f0J2Xbzb/TWv73aP11zR3e9rb/+7wDuKUPjm+x+3776Svz/x7puejbPGSOf8tu/18a4+mVdc0Zvvd7/4ufVXuO8oQuF3t9+kP/8szfCz7458MKo06dy/ubE6qu/LO9Zb0VcX/1z/8aOWP6SyowiFK+qKf/77074tM8ZI5/7lq9Xxaj9gWG9F/+3Xqoodf85sRzEK1xW/+eEvF1+ttuZGurUrvO9VrbuiKNzwqh44ffjT0+3N39bavzTSrV3hQ39y1RVF4QOvzO/Taz6zt53VXhGvzFmd1QZF4T2vrJn41RZ8R51u7dNZ1iTWaiv6geksoepr1VRXIj78daVXEa10676odBgKr7uiO1+5qHQbfHcjwI+779yvLlZ8x8M+XeBPyK6Mut1de0V3Crf/XeAdRSgMsF5QGCBrUBgga1AYIGtQGCBrUBgga1AYIGtQGCBrUBgga1C4QF7e+Wn78cH9Yx8GRAGFC+TTo/vbl1/8cuzDgCigcIlc3/k/v/322AcBcUDhInlycvfYhwCRQOEiuT55eOxDgEigcIl8evTf7/x07IOAOKBwibz84l8f3T/2QUAcULhAPj54uL2mGV4JKFwgT7745bYvzYTWOkBhgKxBYYCsQWGArEFhgKxBYYCsQWGArEFhgKxBYYCsQWGArEFhgKxBYYCsQWGArEFhgKxBYYCs+f+ng3wLkn4oWgAAAABJRU5ErkJggg==" alt="" />
用dbscan算法聚类效果如下。
#用dbscan聚类
p <- ggplot(data,aes(x,y))
library('fpc')
model2 <- dbscan(data,eps=0.2,MinPts=)
p + geom_point(size=2.5, aes(colour=factor(model2$cluster)))+theme(legend.position='top')
aaarticlea/png;base64,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" alt="" />
请注意观察两个算法的不同之处。
1 dbscan是基于密度计算聚类的,会剔除异常(噪声点)。如上图中的类别0,就是dbscan算法聚类出的噪声点(不是核心点且不再核心点的邻域内)。
2 k-means需要指定k值,并且初始聚类中心对聚类结果影响很大。
3 k-means把任何点都归到了某一个类,对异常点比较敏感。
其它的观点。来自《数据挖掘导论》[美]Pang-Ning Tan,Michael Steinbach,Vipin Kumar 著 page355-356
为了简化比较,我们假定对于K均值和DBSCAN都没有距离的限制,并且DBSCAN总是将与若干个核心点相关联的边界点指派到最近的核心点。
1. K均值和DBSCAN都是将每个对象指派到单个簇的划分聚类算法,但是K均值一般聚类所有对象,而DBSCAN丢弃被它识别为噪声的对象。
2. K均值使用簇的基于原型的概念,而DBSCAN使用基于密度的概念。
3. K均值很难处理非球形的簇和不同大小的簇。DBSCAN可以处理不同大小或形状的簇,并且不太受噪声和离群点的影响。当簇具有很不相同的密度时,两种算法的性能都很差。
4. K均值只能用于具有明确定义的质心(比如均值或中位数)的数据。DBSCAN要求密度定义(基于传统的欧几里得密度概念)对于数据是有意义的。
5. K均值可以用于稀疏的高维数据,如文档数据。DBSCAN通常在这类数据上的性能很差,因为对于高维数据,传统的欧几里得密度定义不能很好处理它们。
6. K均值和DBSCAN的最初版本都是针对欧几里得数据设计的,但是它们都被扩展,以便处理其他类型的数据。
7. 基本K均值算法等价于一种统计聚类方法(混合模型),假定所有的簇都来自球形高斯分布,具有不同的均值,但具有相同的协方差矩阵。DBSCAN不对数据的分布做任何假定。
8. K均值DBSCAN和都寻找使用所有属性的簇,即它们都不寻找可能只涉及某个属性子集的簇。
9. K均值可以发现不是明显分离的簇,即便簇有重叠也可以发现,但是DBSCAN会合并有重叠的簇。
10. K均值算法的时间复杂度是O(m),而DBSCAN的时间复杂度是O(m^2),除非用于诸如低维欧几里得数据这样的特殊情况。
11. DBSCAN多次运行产生相同的结果,而K均值通常使用随机初始化质心,不会产生相同的结果。
12. DBSCAN自动地确定簇个数,对于K均值,簇个数需要作为参数指定。然而,DBSCAN必须指定另外两个参数:Eps(邻域半径)和MinPts(最少点数)。
13. K均值聚类可以看作优化问题,即最小化每个点到最近质心的误差平方和,并且可以看作一种统计聚类(混合模型)的特例。DBSCAN不基于任何形式化模型。
所以,不同的数据集和场景,需要运用不同的聚类算法。
下面介绍该算法的工作原理。
aaarticlea/png;base64,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" alt="" />
其中,DBSCAN方法对参数eps和MinPts敏感。
在这个算法框架中,NEps(x, D)表示数据集D中包含在对象x的Eps-邻域范围内的
所有子对象。card(N)表示集合N的基数,即集合N中包含的元素的个数。在簇扩展
过程中采用了栈结构,用于压栈当前对象x的所有邻居对象,再递归地判断栈成员
是否满足核心对象条件,从而决定是否进一步扩展。
后记:
1 关于算法的一般性介绍,可参看百度百科介绍。http://baike.baidu.com/link?url=cnLtGJsF_a4CzmVbAev3nFH75nZUMgwClKv_kk2ZsXuXrP1gvY8eMvY75UDL29AMJFJ2n60xB680PMkjitrG4a
2 按照上述的算法流程,作者写了java代码放入了百度云盘(含上述测试数据),有兴趣的读者请自行下载。http://pan.baidu.com/s/1ntwyXkP
3 参考文献 《针对非均匀数据集的DBSCAN聚类算法研究》 重庆大学硕士学位论文 陈若田 二O 一三年四月 http://pan.baidu.com/s/1mgvKR7U
4 Clustering http://scikit-learn.org/stable/modules/clustering.html#spectral-clustering
5 python DBSCAN源码 http://scikit-learn.org/stable/auto_examples/cluster/plot_dbscan.html#example-cluster-plot-dbscan-py
(完)
基于密度聚类的DBSCAN和kmeans算法比较的更多相关文章
- 聚类算法博客 K-means算法
最近看到一个 blog 感觉超好.记录下.. http://blog.pluskid.org/?p=17
- 机器学习算法总结(五)——聚类算法(K-means,密度聚类,层次聚类)
本文介绍无监督学习算法,无监督学习是在样本的标签未知的情况下,根据样本的内在规律对样本进行分类,常见的无监督学习就是聚类算法. 在监督学习中我们常根据模型的误差来衡量模型的好坏,通过优化损失函数来改善 ...
- 聚类:层次聚类、基于划分的聚类(k-means)、基于密度的聚类、基于模型的聚类
一.层次聚类 1.层次聚类的原理及分类 1)层次法(Hierarchicalmethods)先计算样本之间的距离.每次将距离最近的点合并到同一个类.然后,再计算类与类之间的距离,将距离最近的类合并为一 ...
- DBSCAN密度聚类
1. 密度聚类概念 DBSCAN(Density-Based Spatial Clustering of Applications with Noise,具有噪声的基于密度的聚类方法)是一种很典型的密 ...
- Python机器学习笔记:K-Means算法,DBSCAN算法
K-Means算法 K-Means 算法是无监督的聚类算法,它实现起来比较简单,聚类效果也不错,因此应用很广泛.K-Means 算法有大量的变体,本文就从最传统的K-Means算法学起,在其基础上学习 ...
- 基于ReliefF和K-means算法的医学应用实例
基于ReliefF和K-means算法的医学应用实例 数据挖掘方法的提出,让人们有能力最终认识数据的真正价值,即蕴藏在数据中的信息和知识.数据挖掘 (DataMiriing),指的是从大型数据库或数据 ...
- ML: 聚类算法R包 - 密度聚类
密度聚类 fpc::dbscan fpc::dbscan DBSCAN核心思想:如果一个点,在距它Eps的范围内有不少于MinPts个点,则该点就是核心点.核心和它Eps范围内的邻居形成一个簇.在一个 ...
- 数据挖掘-聚类分析(Python实现K-Means算法)
概念: 聚类分析(cluster analysis ):是一组将研究对象分为相对同质的群组(clusters)的统计分析技术.聚类分析也叫分类分析,或者数值分类.聚类的输入是一组未被标记的样本,聚类根 ...
- 机器学习之K-means算法
前言 以下内容是个人学习之后的感悟,转载请注明出处~ 简介 在之前发表的线性回归.逻辑回归.神经网络.SVM支持向量机等算法都是监督学习算法,需要样本进行训练,且 样本的类别是知 ...
随机推荐
- mysql导入sql文件,乱码,一个例子
服务器centos,其他数据库都是utf8都正常能用 网页正常显示 这次导入一个utf8格式数据库文件 结果网页上乱码 在导出的sql文件中,注释部分有些语句也是被mysql考虑的.导出的sql文件, ...
- 【BZOJ-3156】防御准备 DP + 斜率优化
3156: 防御准备 Time Limit: 10 Sec Memory Limit: 512 MBSubmit: 951 Solved: 446[Submit][Status][Discuss] ...
- hdu 1558 线段相交+并查集
题意:要求相交的线段都要塞进同一个集合里 sol:并查集+判断线段相交即可.n很小所以n^2就可以水过 #include <iostream> #include <cmath> ...
- lua中的table、stack和registery
ok,前面准备给一个dll写wrapper,写了篇日志,看似写的比较明白了,但是其实有很多米有弄明白的.比如PIL中使用的element,key,tname,field这些,还是比较容易混淆的.今天正 ...
- 手机卫士开发记录之json错误
- List多对多的查询应用
/** * @param param * @author mercy 查询主副产品(主副产品是多对多的关系) * @return */ public String queryProductSpecAt ...
- 三角形问题的解决复杂度O(n^3)和O(nlogn)的比较
问题描述: n条棍子组成一个三角形,使得三角形周少最大. 方法一: 暴力解则算法复杂度为O(n^3) #include<stdio.h> const int MAX_N=105 int m ...
- .bash_profile for mac‘ envionment variables
A typical install of OS X won't create a .bash_profile for you. When you want to run functions from ...
- MATLAB cell & struct
CELL cell的每个单元都可以存储任何数据,比如传递函数等.当然,存储矩阵更是没有问题的了.但是用cell数据类型之前,要先初始化. a=cell(n,m) 那么就把a初始化为一个n行m列的空ce ...
- Jboss7.1 加入realm auth认证 bootsfaces 美化的登录页面
jboss-as-7.1.1.Final\standalone\configuration: 1, standalone.xml中 <security-domains>标签里面添加: &l ...