cmt-20230411-1: +other_models:knn+svm

other_models/.gitignore

@@ -0,0 +1,4 @@
+.idea
+**/*.txt
+**/*.png
+**/*.jpg

other_models/kmeans_gray/batch_plot.py

@@ -0,0 +1,80 @@
+from travel_path import *
+import pylab as pl
+
+mark = ['ob', 'or', 'og', 'ok', '^r', '+r', 'sr', 'dr', '<r', 'pr']
+
+files = travel_txt('cluster_freqXstd/')
+files = files.reshape(20,1)
+files1 = np.copy(files[6:16])
+files2 = np.row_stack((files[16:],files[:6]))
+
+xlines = np.genfromtxt('Mild Outlier.txt')
+xlines = xlines.reshape(20,1)
+xlines1 = np.copy(xlines[6:16])
+xlines2 = np.row_stack((xlines[16:],xlines[:6]))
+
+pl.figure(figsize=(10,18))
+for i in range(files1.shape[0]):
+    data = np.genfromtxt(files1[i,0])
+
+    SpacePoint = files1[i,0].split(' ')[-1]
+    SpacePoint = SpacePoint.split('.')[0] + '.' + SpacePoint.split('.')[1]
+    print(SpacePoint)
+
+    data_tmp = data[:,0]
+    max = np.max(data_tmp, axis=0)
+    min = np.min(data_tmp, axis=0)
+
+    pl.subplot(5,2,i+1)
+    for j in range(data.shape[0]):
+        if data[j,2] == 1:
+            pl.plot(data[j, 0], data[j, 1], mark[np.int(data[j, 2])], label='Relatively Abnormal Signal')
+        else:
+            pl.plot(data[j, 0], data[j, 1], mark[np.int(data[j, 2])])
+    pl.legend()
+    pl.xlabel("timeStd / V")
+    pl.ylabel("freqCorr")
+    pl.title("2-Clustering Result of freqXstd of Point at " + SpacePoint)
+
+    # x = np.arange(min, max, (max-min)/data.shape[0])
+    # if x.shape[0] > 1728:
+    #     x = x[:1728]
+    #
+    # pl.axvline(xlines1[i], color='k', linewidth=0.5)
+    # yline = 0.55*np.ones(data.shape[0])
+    # pl.plot(x, yline, color='k', linewidth=0.5)
+pl.tight_layout()
+pl.savefig('cluster_freqXstd/fig1.png')
+
+pl.figure(figsize=(10,18))
+for i in range(files2.shape[0]):
+    data = np.genfromtxt(files2[i,0])
+
+    SpacePoint = files2[i,0].split(' ')[-1]
+    SpacePoint = SpacePoint.split('.')[0] + '.' + SpacePoint.split('.')[1]
+    print(SpacePoint)
+
+    data_tmp = data[:,0]
+    max = np.max(data_tmp, axis=0)
+    min = np.min(data_tmp, axis=0)
+
+    pl.subplot(5,2,i+1)
+    for j in range(data.shape[0]):
+        if data[j,2] == 1:
+            pl.plot(data[j, 0], data[j, 1], mark[np.int(data[j, 2])], label='Relatively Abnormal Signal')
+        else:
+            pl.plot(data[j, 0], data[j, 1], mark[np.int(data[j, 2])])
+    pl.legend()
+    pl.xlabel("timeStd / V")
+    pl.ylabel("freqCorr")
+    pl.title("2-Clustering Result of freqXstd of Point at " + SpacePoint)
+
+    # x = np.arange(min, max, (max-min)/data.shape[0])
+    # if x.shape[0] > 1728:
+    #     x = x[:1728]
+
+    # pl.axvline(xlines2[i], color='k', linewidth=0.5)
+    # yline = 0.55*np.ones(data.shape[0])
+    # pl.plot(x, yline, color='k', linewidth=0.5)
+pl.tight_layout()
+pl.savefig('cluster_freqXstd/fig2.png')

other_models/kmeans_gray/cluster_freqStd/2clusters/scatter.py

@@ -0,0 +1,11 @@
+import pylab as pl
+import numpy as np
+
+mark = ['ob', 'or', 'og', 'ok', '^r', '+r', 'sr', 'dr', '<r', 'pr']
+
+data = np.genfromtxt("2 clusters of point at 3.1km.txt")
+for i in range(data.shape[0]):
+    pl.plot(data[i,0],data[i,1], mark[np.int(data[i,2])])
+pl.xlabel("timeStd / V")
+pl.ylabel("freqCorr")
+pl.show()

other_models/kmeans_gray/cluster_freqStd/3clusters/scatter.py

@@ -0,0 +1,11 @@
+import pylab as pl
+import numpy as np
+
+mark = ['og', 'ob', 'or', 'ok', '^r', '+r', 'sr', 'dr', '<r', 'pr']
+
+data = np.genfromtxt("3 clusters of point at 3.6km.txt")
+for i in range(data.shape[0]):
+    pl.plot(data[i,0],data[i,1], mark[np.int(data[i,2])])
+pl.xlabel("timeStd / V")
+pl.ylabel("freqCorr")
+pl.show()

other_models/kmeans_gray/cluster_freqStd/4clusters/scatter.py

@@ -0,0 +1,11 @@
+import pylab as pl
+import numpy as np
+
+mark = ['ok', 'og', 'ob', 'or', '^r', '+r', 'sr', 'dr', '<r', 'pr']
+
+data = np.genfromtxt("4 clusters of point at 3.1km.txt")
+for i in range(data.shape[0]):
+    pl.plot(data[i,0],data[i,1], mark[np.int(data[i,2])])
+pl.xlabel("timeStd / V")
+pl.ylabel("freqCorr")
+pl.show()

other_models/kmeans_gray/cluster_freqXstd/scatter.py

@@ -0,0 +1,11 @@
+import pylab as pl
+import numpy as np
+
+mark = ['ob', 'or', 'og', 'ok', '^r', '+r', 'sr', 'dr', '<r', 'pr']
+
+data = np.genfromtxt("2 clusters of point at 3.6km.txt")
+for i in range(data.shape[0]):
+    pl.plot(data[i,0],data[i,1], mark[np.int(data[i,2])])
+pl.xlabel("timeStd / V")
+pl.ylabel("freqCorr")
+pl.show()

other_models/kmeans_gray/freqCorr/plot.py

@@ -0,0 +1,14 @@
+import pylab as pl
+import numpy as np
+
+data = np.genfromtxt('freq corr of point at 3.6km.txt')
+
+# x = np.linspace(0,1728,1728)
+pl.xlim(0, data.shape[0])
+# pl.ylim(0.6)
+pl.plot(data)
+
+# x = 0.65*np.ones(data.shape[0])
+# pl.plot(x, color='red')
+
+pl.show()

other_models/kmeans_gray/freqCorr/read.py

@@ -0,0 +1,4 @@
+import numpy as np
+
+data = np.genfromtxt("freq corr of point at 3.1km.txt")
+print(data.shape)

other_models/kmeans_gray/getFreqStd.py

@@ -0,0 +1,24 @@
+from travel_path import *
+
+freqCorr_files = travel_txt("freqCorr/")
+timeStd_files = travel_txt("timeStd/")
+
+for i in range(freqCorr_files.shape[0]):
+    SpacePoint = freqCorr_files[i].split(' ')[-1]
+    SpacePoint = SpacePoint.split('.')[0] + '.' + SpacePoint.split('.')[1] + '.' + SpacePoint.split('.')[-1]
+    SpacePoint1 = timeStd_files[i].split('_')[-1]
+    SpacePoint1 = SpacePoint1.split('.')[0] + '.' + SpacePoint1.split('.')[1] + '.' + SpacePoint.split('.')[-1]
+    if SpacePoint == SpacePoint1:
+        print("Bingo!")
+    print(SpacePoint)
+
+    data_freCorr = np.genfromtxt(freqCorr_files[i])
+    data_timeStd = np.genfromtxt(timeStd_files[i])
+    data_timeStd = np.transpose(data_timeStd)
+    data_timeStd = data_timeStd.reshape(1728, )
+    # Normalization
+    data_timeStd = data_timeStd/np.max(data_timeStd)
+    data_freqStd = np.column_stack((data_timeStd, data_freCorr))
+    print(data_freqStd.shape)
+
+    np.savetxt("freqStd/freqStd of point at " + SpacePoint, data_freqStd)

other_models/kmeans_gray/kmeans_2clusters.py

@@ -0,0 +1,52 @@
+# -*- coding: utf-8 -*-
+import cv2
+from travel_path import *
+import math
+
+# calculate Euclidean distance
+def euclDistance(vector1, vector2):
+    return math.sqrt(sum(pow(vector2 - vector1, 2)))
+
+freqStd_files = travel_txt("freqStd/")
+
+# define criteria, number of clusters(K) and apply kmeans()
+criteria = (cv2.TERM_CRITERIA_EPS + cv2.TERM_CRITERIA_MAX_ITER, 10, 1.0)
+
+for i in range(freqStd_files.shape[0]):
+    SpacePoint = freqStd_files[i].split(' ')[-1]
+    SpacePoint = SpacePoint.split('.')[0] + '.' + SpacePoint.split('.')[1] + '.' + SpacePoint.split('.')[-1]
+    print(SpacePoint)
+
+    data = np.genfromtxt(freqStd_files[i])
+    data_1 = data
+    data_1 = np.float32(data_1)
+    # k聚类
+    ret, label, center = cv2.kmeans(data_1, 2, None, criteria, 10, cv2.KMEANS_RANDOM_CENTERS)
+    print(label.shape)
+
+    suffix = np.array([[0],[1]])
+    center = np.column_stack((center, suffix))
+    print(center)
+    for j in range(2):
+        for k in range(j+1,2):
+            vec0 = np.array([0,0])
+            vec1 = center[j,:2]
+            vec2 = center[k,:2]
+            if euclDistance(vec1,vec0) > euclDistance(vec2,vec0):
+                tmp1 = np.copy(center[k,:])
+                tmp2 = np.copy(center[j,:])
+                center[j,:] = tmp1
+                center[k,:] = tmp2
+    print(center)
+
+    print(label)
+    label1 = np.copy(label)
+    for j in range(label.shape[0]):
+        for k in range(2):
+            if label[j,0] == center[k,2]:
+                label1[j,0] = np.copy(k)
+    print(label1)
+
+    data_new = np.column_stack((data, label1))
+    print(data_new.shape)
+    np.savetxt("cluster_freqStd/2clusters/2 clusters of point at " + SpacePoint, data_new)

other_models/kmeans_gray/kmeans_2clusters_multi.py

@@ -0,0 +1,50 @@
+# -*- coding: utf-8 -*-
+import cv2
+from travel_path import *
+
+freqStd_files = travel_txt("freqStd/")
+
+# define criteria, number of clusters(K) and apply kmeans()
+criteria = (cv2.TERM_CRITERIA_EPS + cv2.TERM_CRITERIA_MAX_ITER, 10, 1.0)
+
+for i in range(freqStd_files.shape[0]):
+    SpacePoint = freqStd_files[i].split(' ')[-1]
+    SpacePoint = SpacePoint.split('.')[0] + '.' + SpacePoint.split('.')[1] + '.' + SpacePoint.split('.')[-1]
+    print(SpacePoint)
+
+    data = np.genfromtxt(freqStd_files[i])
+    data_multi = []
+    for j in range(data.shape[0]):
+        if data_multi == []:
+            data_multi = data[j,0] * data[j,1]
+        else:
+            data_tmp = data[j,0] * data[j,1]
+            data_multi = np.column_stack((data_multi, data_tmp))
+    data_multi = np.float32(data_multi)
+    # k聚类
+    ret, label, center = cv2.kmeans(data_multi, 2, None, criteria, 10, cv2.KMEANS_RANDOM_CENTERS)
+    print(label.shape)
+
+    suffix = np.array([[0],[1]])
+    center = np.column_stack((center, suffix))
+    print(center)
+    for j in range(2):
+        for k in range(j+1,2):
+            if center[j,0] > center[k,0]:
+                tmp1 = np.copy(center[k,:])
+                tmp2 = np.copy(center[j,:])
+                center[j,:] = tmp1
+                center[k,:] = tmp2
+    print(center)
+
+    print(label)
+    label1 = np.copy(label)
+    for j in range(label.shape[0]):
+        for k in range(2):
+            if label[j,0] == center[k,1]:
+                label1[j,0] = np.copy(k)
+    print(label1)
+
+    data_new = np.column_stack((data, label1))
+    print(data_new.shape)
+    np.savetxt("cluster_freqXstd/2 clusters of point at " + SpacePoint, data_new)

other_models/kmeans_gray/kmeans_3clusters.py

@@ -0,0 +1,52 @@
+# -*- coding: utf-8 -*-
+import cv2
+from travel_path import *
+import math
+
+# calculate Euclidean distance
+def euclDistance(vector1, vector2):
+    return math.sqrt(sum(pow(vector2 - vector1, 2)))
+
+freqStd_files = travel_txt("freqStd/")
+
+# define criteria, number of clusters(K) and apply kmeans()
+criteria = (cv2.TERM_CRITERIA_EPS + cv2.TERM_CRITERIA_MAX_ITER, 10, 1.0)
+
+for i in range(freqStd_files.shape[0]):
+    SpacePoint = freqStd_files[i].split(' ')[-1]
+    SpacePoint = SpacePoint.split('.')[0] + '.' + SpacePoint.split('.')[1] + '.' + SpacePoint.split('.')[-1]
+    print(SpacePoint)
+
+    data = np.genfromtxt(freqStd_files[i])
+    data_1 = data
+    data_1 = np.float32(data_1)
+    # k聚类
+    ret, label, center = cv2.kmeans(data_1, 3, None, criteria, 10, cv2.KMEANS_RANDOM_CENTERS)
+    print(label.shape)
+
+    suffix = np.array([[0],[1],[2]])
+    center = np.column_stack((center, suffix))
+    print(center)
+    for j in range(3):
+        for k in range(j+1,3):
+            vec0 = np.array([0, 0])
+            vec1 = center[j, :2]
+            vec2 = center[k, :2]
+            if euclDistance(vec1, vec0) > euclDistance(vec2, vec0):
+                tmp1 = np.copy(center[k,:])
+                tmp2 = np.copy(center[j,:])
+                center[j,:] = tmp1
+                center[k,:] = tmp2
+    print(center)
+
+    print(label)
+    label1 = np.copy(label)
+    for j in range(label.shape[0]):
+        for k in range(3):
+            if label[j,0] == center[k,2]:
+                label1[j,0] = np.copy(k)
+    print(label1)
+
+    data_new = np.column_stack((data, label1))
+    print(data_new.shape)
+    np.savetxt("cluster_freqStd/3clusters/3 clusters of point at " + SpacePoint, data_new)

other_models/kmeans_gray/kmeans_4clusters.py

@@ -0,0 +1,52 @@
+# -*- coding: utf-8 -*-
+import cv2
+from travel_path import *
+import math
+
+# calculate Euclidean distance
+def euclDistance(vector1, vector2):
+    return math.sqrt(sum(pow(vector2 - vector1, 2)))
+
+freqStd_files = travel_txt("freqStd/")
+
+# define criteria, number of clusters(K) and apply kmeans()
+criteria = (cv2.TERM_CRITERIA_EPS + cv2.TERM_CRITERIA_MAX_ITER, 10, 1.0)
+
+for i in range(freqStd_files.shape[0]):
+    SpacePoint = freqStd_files[i].split(' ')[-1]
+    SpacePoint = SpacePoint.split('.')[0] + '.' + SpacePoint.split('.')[1] + '.' + SpacePoint.split('.')[-1]
+    print(SpacePoint)
+
+    data = np.genfromtxt(freqStd_files[i])
+    data_1 = data
+    data_1 = np.float32(data_1)
+    # k聚类
+    ret, label, center = cv2.kmeans(data_1, 4, None, criteria, 10, cv2.KMEANS_RANDOM_CENTERS)
+    print(label.shape)
+
+    suffix = np.array([[0],[1],[2],[3]])
+    center = np.column_stack((center, suffix))
+    print(center)
+    for j in range(4):
+        for k in range(j+1,4):
+            vec0 = np.array([0, 0])
+            vec1 = center[j, :2]
+            vec2 = center[k, :2]
+            if euclDistance(vec1, vec0) > euclDistance(vec2, vec0):
+                tmp1 = np.copy(center[k,:])
+                tmp2 = np.copy(center[j,:])
+                center[j,:] = tmp1
+                center[k,:] = tmp2
+    print(center)
+
+    print(label)
+    label1 = np.copy(label)
+    for j in range(label.shape[0]):
+        for k in range(4):
+            if label[j,0] == center[k,2]:
+                label1[j,0] = np.copy(k)
+    print(label1)
+
+    data_new = np.column_stack((data, label1))
+    print(data_new.shape)
+    np.savetxt("cluster_freqStd/4clusters/4 clusters of point at " + SpacePoint, data_new)