cfs.rar_CFS聚类_announced1vd_cfs聚类算法_python_python 聚类

#!/usr/bin/env python #-*-coding: utf-8 -*- import os import math import sys import string import time import re DIM_PARAM_NUMBER = 12 pre_cuid = '' pre_time = '' loc_list = [] dc = 1000 k = 5 def eucdist(a, b): sum = 0.0 for i in range(1, len(a)): sum += math.pow(a[i] - b[i], 2) sum = math.sqrt(sum) return sum def dist(c): D = [] for i in c: td = [] for j in c: di = eucdist(i, j) td.append(di) D.append(td) return def lfind(d, dc, count): index = [] c = 0 for i in d[count]: if i <= dc: index.append(c) c += 1 return index def gridSplit(data, eps): dim = len(data[0]) grid = [] for i in range(1, dim): g_i = [] max_i = -9999 for j in data: if j[i] > max_i: max_i = j[i] for j in data: g_i.append(int((j[i] - max_i)/eps)) grid.append(g_i) return grid def gm(grid, i): m = len(grid) s = "" for j in range(0, m): s += str(grid[j][i]) + "_" s = s[:-1] return s def gridMap(grid): grid_map = {} n = len(grid[0]) for i in range(0, n): s = gm(grid, i) if s not in grid_map: grid_map[s] = [] grid_map[s].append(i) return grid_map def ccand(grid_map, cand, gd_array, i): if i == len(gd_array): s = "_".join(gd_array) if s in grid_map: cand += grid_map[s] return gd_array[i] = str(int(gd_array[i]) + 1) ccand(grid_map, cand, gd_array, i+1) gd_array[i] = str(int(gd_array[i]) - 2) ccand(grid_map, cand, gd_array, i+1) gd_array[i] = str(int(gd_array[i]) + 1) ccand(grid_map, cand, gd_array, i+1) def gfind(data, grid, grid_map, eps, count): gd = gm(grid, count) gd_array = gd.split("_") cand = [] ccand(grid_map, cand, gd_array, 0) index = [] for i in cand: s = "" if i <= count: s = str(i) + "_" + str(count) else: s = str(count) + "_" + str(i) di = eucdist(data[i], data[count]) if di < eps: index.append(i) return index def cfs(data, cuid): n = len(data) sys.__stderr__.write("%s\n" % (n)) if n < k: return {} if not re.match(r".+\|.+", cuid): return dbscan_gps(data, k, dc, cuid) visited = [0]*n p = [] grid = gridSplit(data, dc) grid_map = gridMap(grid) for i in range(0,len(data)): neighbor = gfind(data, grid, grid_map, dc, i) p.append(len(neighbor))