def clusterCore(channelData1, covMatrixList1, channelData2, centroids,
centroidUList, type):
newChannelData1 = []
newChannelData2 = []
newDimension = np.shape(centroidUList[0])[1]
p = np.shape(channelData1)[0]
if type == "C":
# 计算信道相关系数矩阵并输出，然后放到一个矩阵中
allCovMatrix1 = tools.matrixListToMatrix(covMatrixList1)
# 确定每个数据分别属于哪个簇
clusterAssment = kmeans.getClusterAssment(allCovMatrix1, centroids)
# 变换域
for i in range(p):
newChannelData1.append(
np.dot(channelData1[i],
centroidUList[(int)(clusterAssment[i, 0].real)]))
newChannelData2.append(
np.dot(channelData2[i],
centroidUList[(int)(clusterAssment[i, 0].real)]))
if type == "U":
informations, SigmaList, UList = tools.getInformations(covMatrixList1)
allU = tools.matrixListToMatrix_U(UList)
weights = tools.matrixListToMatrix_U(SigmaList)
# 确定每个数据分别属于哪个簇
clusterAssment = kmeans.getClusterAssment_U(allU, weights, centroids,
newDimension)
# 变换域
for i in range(p):
newChannelData1.append(
np.dot(channelData1[i],
centroidUList[(int)(clusterAssment[i, 0].real)]))
newChannelData2.append(
np.dot(channelData2[i],
centroidUList[(int)(clusterAssment[i, 0].real)]))
if type == "general":
newChannelData1 = pca.pca_general(channelData1, newDimension)
newChannelData2 = pca.pca_general(channelData2, newDimension)
if type == "none":
newChannelData1 = channelData1
newChannelData2 = channelData2
allNewCorr = []
for i in range(p):
for j in range(newDimension):
cowCor = np.corrcoef(newChannelData1[i][:, j],
newChannelData2[i][:, j])
if i == 0:
allNewCorr.append(cowCor[0, 1])
else:
allNewCorr[j] += cowCor[0, 1]
for i in range(newDimension):
allNewCorr[i] = abs(allNewCorr[i] / (np.shape(channelData1)[0]))
path = u'/Users/jinruimeng/Downloads/keyan/'
nowTime = time.strftime("%Y-%m-%d.%H.%M.%S", time.localtime(time.time()))
pathSuffix = type + u'_' + nowTime
newChannelData1Path = path + "clusterAddNoise_newChannelData1_" + pathSuffix
newChannelData2Path = path + "clusterAddNoise_newChannelData2_" + pathSuffix
readAndWriteDataSet.write(newChannelData1, newChannelData1Path, ".xlsx")
readAndWriteDataSet.write(newChannelData2, newChannelData2Path, ".xlsx")
return allNewCorr
# 路径配置
path = u'/Users/jinruimeng/Downloads/keyan/'
# path = u'E:\\workspace\\keyan\\'
# 信噪比的上下限
low = -10
high = 5
step = 5
suffix = str(low) + u'to' + str(high) + u'dB'
time1 = ((int)((high - low) / step + 1))
# 读入信道数据
channelDataPath = path + u'channelDataP.xlsx'
channelDataAll = readAndWriteDataSet.excelToMatrixList(channelDataPath)
p = np.shape(channelDataAll)[0] # 页数
# 添加噪声
channelDataAll1 = []
channelDataAll2 = []
for h in range(time1):
SNR = low + h * step
for i in range(p):
noise1, noise2, npower = wgn_abs(channelDataAll[i], SNR)
channelDataAll1.append(channelDataAll[i] + noise1)
channelDataAll2.append(channelDataAll[i] + noise2)
outChannelAll1ListPath = path + "outChannelAll1List_" + suffix
outChannelAll2ListPath = path + "outChannelAll2List_" + suffix
readAndWriteDataSet.write(channelDataAll1, outChannelAll1ListPath, ".xlsx")
readAndWriteDataSet.write(channelDataAll2, outChannelAll2ListPath, ".xlsx")
totalOldCorr, totalPcaCorr, totalNewCCorr, totalNewUCorr = cluster(
a, schedule, channelDataAll1, channelDataAll2, allCentroidsC,
allCentroidUList, allCentroidsU, allCentroidUList2)
ps.close()
ps.join()
# 输出相关系数
path = u'/Users/jinruimeng/Downloads/keyan/'
nowTime = time.strftime("%Y-%m-%d.%H.%M.%S", time.localtime(time.time()))
pathSuffix = str(SNR) + u'_' + nowTime
outOldCorrPath = path + "clusterAddNoise_outOldCorr_" + pathSuffix
outPcaCorrPath = path + "clusterAddNoise_outPcaCorr_" + pathSuffix
outNewCCorrPath = path + "clusterAddNoise_outNewCCorr_" + pathSuffix
outNewUCorrPath = path + "clusterAddNoise_outNewUCorr_" + pathSuffix
readAndWriteDataSet.write(totalOldCorr, outOldCorrPath, ".xlsx")
readAndWriteDataSet.write(totalPcaCorr, outPcaCorrPath, ".xlsx")
readAndWriteDataSet.write(totalNewCCorr, outNewCCorrPath, ".xlsx")
readAndWriteDataSet.write(totalNewUCorr, outNewUCorrPath, ".xlsx")
for i in range(np.shape(totalPcaCorr)[0]):
plt.xlabel(u'分量序号')
X = range(low, high + 1, step)
plt.ylabel(u'相关系数')
plt.plot(X, totalOldCorr[i], 'k-s', label=u'预处理前')
plt.plot(X, totalPcaCorr[i], 'g-^', label=u'无交互PCA')
plt.plot(X, totalNewCCorr[i], 'r-v', label=u'聚类协方差矩阵')
plt.plot(X, totalNewUCorr[i], 'b-o', label=u'聚类变换矩阵')
plt.legend(bbox_to_anchor=(0., 1.02, 1., .102),
loc=0,
ncol=4,
ps.close()
ps.join()
# 输出不一致率
total_inconsistency_rate_old_array = readAndWriteDataSet.listToArray(total_inconsistency_rate_old)
total_inconsistency_rate_new_noCom_array = readAndWriteDataSet.listToArray(total_inconsistency_rate_new_noCom)
total_inconsistency_rate_new_array = readAndWriteDataSet.listToArray(total_inconsistency_rate_new)
path = u'/Users/jinruimeng/Downloads/keyan/'
nowTime = time.strftime("%Y-%m-%d.%H.%M.%S", time.localtime(time.time()))
pathSuffix = u'SNR：' + str(low) + u'to' + str(high) + u'_' + u'bit：' + str(low2) + u'to' + str(high2) + u'_' + nowTime
outOldInconsistencyPath = path + "clusterAddNoise_outOldInconsistency_" + pathSuffix
outNewInconsistencyNoComPath = path + "clusterAddNoise_outNewInconsistencyNoCom_" + pathSuffix
outNewInconsistencyPath = path + "clusterAddNoise_outNewInconsistency_" + pathSuffix
readAndWriteDataSet.write(total_inconsistency_rate_old_array, outOldInconsistencyPath, ".xlsx")
readAndWriteDataSet.write(total_inconsistency_rate_new_noCom_array, outNewInconsistencyNoComPath, ".xlsx")
readAndWriteDataSet.write(total_inconsistency_rate_new_array, outNewInconsistencyPath, ".xlsx")
# fig = plt.figure()
# ax = fig.gca(projection='3d')
# ax.set_xlabel(u'信噪比（dB）')
# ax.set_ylabel(u'量化比特数')
# ax.set_zlabel(u'密钥不一致率')
#
# x = []
# y = []
# z1 = []
# z2 = []
# for i in range(low, high + 1, step):
# for j in range(low2, high2 + 1, step2):
newSNRsArray[i, j] = newSNRsArray[i, j] / ((1 << a) * p)
newSNRsArray[i, j] = 10 * np.log10(newSNRsArray[i, j])
newSNRsList.append(newSNRsArray)
readAndWriteDataSet.writeKey(path, key_newPca1, key_newPca2, SNR, u'pca')
readAndWriteDataSet.writeKey(path, key_newC1, key_newC2, SNR, u'C')
readAndWriteDataSet.writeKey(path, key_newU1, key_newU2, SNR, u'U')
readAndWriteDataSet.writeKey(path, key_newWt1, key_newWt2, SNR, u'wt')
# 计算信息熵
total_key_pca = u''
total_key_C = u''
total_key_U = u''
total_key_wt = u''
for i in range(np.shape(key_newPca1)[0]):
total_key_pca = total_key_pca + key_newPca1[i] + key_newPca2[i]
total_key_C = total_key_C + key_newC1[i] + key_newC2[i]
total_key_U = total_key_U + key_newU1[i] + key_newU2[i]
total_key_wt = total_key_wt + key_newWt1[i] + key_newWt2[i]
ents[0, h] = quantification.calc_shannon_ent(total_key_pca)
ents[1, h] = quantification.calc_shannon_ent(total_key_C)
ents[2, h] = quantification.calc_shannon_ent(total_key_U)
ents[3, h] = quantification.calc_shannon_ent(total_key_wt)
readAndWriteDataSet.write(newSNRsList, path + u'newSNR_' + suffix)
readAndWriteDataSet.write(lengths, path + u'length_' + suffix)
readAndWriteDataSet.write(errorNums, path + u'errorNum_' + suffix)
readAndWriteDataSet.write(ents, path + u'ent_' + suffix)
print("主进程结束！")
def getCentroidsCore(path,
suffix,
channelData,
covMatrixList,
informations,
SigmaList,
UList,
g,
k,
iRate,
type="C"):
nowTime = time.strftime("%Y-%m-%d.%H.%M.%S", time.localtime(time.time()))
pathSuffix = type + "_" + str(g) + "_" + nowTime
outOldCovMatrixListPath = path + "getCentroids_outOldCovMatrixList_" + pathSuffix
outCentroidListPath = path + "getCentroids_outCentroidList_" + pathSuffix
outClusterAssmentPath = path + "getCentroids_outClusterAssment_" + pathSuffix
outNewChannelDataPath = path + "getCentroids_outNewChannelData_" + pathSuffix
outNewCovMatrixListPath = path + "getCentroids_outNewCovMatrixList_" + pathSuffix
ratesPath = path + "getCentroids_rates_" + pathSuffix
UTsPath = path + "getCentroids_UTs_" + pathSuffix
clusterAssmentList = []
newChannelData = []
newCovMatrixList = []
UTs = []
rates = []
centroidList = []
if type == u'C':
allCovMatrix = tools.matrixListToMatrix(covMatrixList)
# 对协方差进行聚类
centroids, clusterAssment = kmeans.KMeansOushi(allCovMatrix, k)
clusterAssmentList.append(clusterAssment)
centroidList = tools.matrixToMatrixList(centroids)
# 分析PCA效果
# newChannelData, newCovMatrixList, UTs, rates = pca.pca(channelData, informations, centroidList, clusterAssment, iRate)
if type == u'U':
allU = tools.matrixListToMatrix_U(UList)
weights = tools.matrixListToMatrix_U(SigmaList)
# 对协方差进行聚类
centroids, clusterAssment = kmeans.KMeansOushi_U(
allU, k, weights, iRate)
clusterAssmentList.append(clusterAssment)
centroidList = tools.matrixToMatrixList_U(centroids)
# 分析PCA效果
# newChannelData, newCovMatrixList, UTs, rates = pca.pca_U(channelData, informations, centroidList, clusterAssment, iRate)
# 输出结果
# 输出聚类结果
# readAndWriteDataSet.write(clusterAssmentList, outClusterAssmentPath, suffix)
# 协方差矩阵太大了，先不输出
# readAndWriteDataSet.write(covMatrixList, outOldCovMatrixListPath, suffix)
# 聚类中心太大了，先不输出
readAndWriteDataSet.write(centroidList, outCentroidListPath, suffix)
def cluster(schedule, path, suffix, channelData, g, iRate):
if iRate > np.shape(channelData)[1]:
print(u'降维后维度不能大于样本原有的维度！')
return
if iRate <= 0:
print(u'降维后维度不能小于1！')
return
schedule[1] += 1
tmpSchedule = schedule[1]
print(u'共' + str(schedule[0]) + u'部分，' + u'第' + str(tmpSchedule) +
u'部分开始！')
pathSuffix = "C" + "_" + str(g) + "_"
centroidListPath = path + "getCentroids_outCentroidList_" + pathSuffix
nowTime = time.strftime("%Y-%m-%d.%H.%M.%S", time.localtime(time.time()))
pathSuffix = pathSuffix + str(nowTime)
outOldCovMatrixListPath = path + "cluster_outOldCovMatrixList_" + pathSuffix
outClusterAssmentPath = path + "cluster_outClusterAssment_" + pathSuffix
outNewChannelDataPath = path + "cluster_outNewChannelData_" + pathSuffix
outNewCovMatrixsPath = path + "cluster_outNewCovMatrixList_" + pathSuffix
ratesPath = path + "cluster_rates_" + pathSuffix
UTsPath = path + "cluster_UTs_" + pathSuffix
# 读入聚类中心信息
# 合并多个文件
centroidList = []
for root, dirs, files in os.walk(path, topdown=True):
for file in files:
file = os.path.join(root, file)
if centroidListPath in file:
centroidListTmp = readAndWriteDataSet.excelToMatrixList(file)
for centroid in centroidListTmp:
centroidList.append(centroid)
break
centroids = tools.matrixListToMatrix(centroidList)
# 计算信道相关系数矩阵并输出，然后放到一个矩阵中
covMatrixList = tools.getCovMatrixList(channelData)
allCovMatrix = tools.matrixListToMatrix(covMatrixList)
# 确定每个数据分别属于哪个簇
clusterAssment = kmeans.getClusterAssment(allCovMatrix, centroids)
clusterAssmentList = []
clusterAssmentList.append(clusterAssment)
# 分析PCA效果
newChannelData, newCovMatrixList, UTs, rates = pca.pca(
channelData, covMatrixList, centroidList, clusterAssment, iRate)
# 输出结果
# 输出聚类结果
readAndWriteDataSet.write(clusterAssmentList, outClusterAssmentPath,
suffix)
# 协方差矩阵太大了，先不输出
# readAndWriteDataSet.write(covMatrixList, outOldCovMatrixListPath, suffix)
# 输出PCA结果
readAndWriteDataSet.write(newChannelData, outNewChannelDataPath, suffix)
readAndWriteDataSet.write(newCovMatrixList, outNewCovMatrixsPath, suffix)
readAndWriteDataSet.write(UTs, UTsPath, suffix)
readAndWriteDataSet.write(rates, ratesPath, suffix)
# 显示进度
print(u'共' + str(schedule[0]) + u'部分，' + u'第' + str(tmpSchedule) +
u'部分完成，' + u'已完成' + str(schedule[1]) + u'部分，' + u'完成度：' + '%.2f%%' %
(schedule[1] / schedule[0] * 100) + u'！')