######If you want to select particular moments in the day timeShift = -8 # San Francisco data = ld.dataSelectTime(data, 0, 24, [0, 1, 2, 5, 6], timeShift) ####### Modulo 24 hours # data = clus.temporalAround24removed(data) data = clus.temporalAround7Daysremoved(data) #######Remove time # data=clus.removeTime(data) kk = cluster.KMeans(n_clusters=5, random_state=0) out = kk.fit(data) clus.plotCluster(data, out.labels_, 5) print("Cluster center " + str(out.cluster_centers_)) ld.fromDataClusterToCsv(data, out.labels_, "WorkingDays") ###### Select a center and get the 80% closest values to this center removed to far points partialdata = clus.getThe80percentClosestToACenter(data, out, 0) print(partialdata) ###### Second cluster on the partial data """ kk2 = cluster.KMeans(n_clusters=3,random_state=0) out2 = kk2.fit(partialdata) print(out2.cluster_centers_) clus.plotCluster(partialdata,out2.labels_,3)
data = ld.dataSelectTime(data,10,16,[0,1,2,3,4],1) data = clus.temporalAround24removed(data) data=clus.removeTime(data) kk = cluster.KMeans(n_clusters=10,random_state=0) out = kk.fit(data) clus.plotCluster(data,out.labels_,10) print("Cluster center " + str(out.cluster_centers_)) ld.fromDataClusterToCsv(data,out.labels_,"Work") partialdata = clus.getThe80percentClosestToACenter(data,out,0) print(partialdata) kk2 = cluster.KMeans(n_clusters=3,random_state=0) out2 = kk2.fit(partialdata) print(out2.cluster_centers_) clus.plotCluster(partialdata,out2.labels_,3) ld.fromDataClusterToCsv(partialdata,out2.labels_,"Home") #x,y,t = ld.getXYTForClassI(partialdata,1)