def project_vectors(origForeignVecFile,
origEnVecFile,
subsetEnVecFile,
subsetForeignVecFile,
outputEnFile,
outputForeignFile,
NUMCC=40):
'''
将词典的向量输入到CCA中，生成投影向量，再生成双语向量
:param origForeignVecFile: 外语向量矩阵
:param origEnVecFile: 英语向量矩阵
:param subsetEnVecFile: 词典中的英语向量矩阵
:param subsetForeignVecFile: 词典中的外语向量矩阵
:param outputEnFile: 重新获得的英语词向量
:param outputForeignFile: 重新获得的外语词向量
:param truncRatio: 模型的训练系数
'''
'''数据读入，处理掉开头的英文单词，只保留词向量'''
tmp = np.loadtxt(origEnVecFile, dtype=np.str, delimiter=' ')
origEnVecs = tmp[:, 1:].astype(np.float)
tmp2 = np.loadtxt(origForeignVecFile, dtype=np.str, delimiter=' ')
origForeignVecs = tmp2[:, 1:].astype(np.float)
tmp3 = np.loadtxt(subsetEnVecFile, dtype=np.str, delimiter=' ')
subsetEnVecs = tmp3[:, 1:].astype(np.float)
tmp4 = np.loadtxt(subsetForeignVecFile, dtype=np.str, delimiter=' ')
subsetForeignVecs = tmp4[:, 1:].astype(np.float)
'''预处理，使每行正则化'''
#origEnVecs=preprocessing.normalize(origEnVecs)
#origForeignVecs=preprocessing.normalize(origForeignVecs)
subsetEnVecs = preprocessing.normalize(subsetEnVecs)
subsetForeignVecs = preprocessing.normalize(subsetForeignVecs)
'''训练CCA'''
'''
num = [NUMCC]
regs = [1e-1]
cca = rcca.CCACrossValidate(regs=regs,numCCs=num,kernelcca=False,cutoff=0.1)
cca.train([subsetEnVecs, subsetForeignVecs])
'''
cca = CCA(n_components=NUMCC)
cca.fit(subsetEnVecs, subsetForeignVecs)
print cca.get_params()
X_c, Y_c = cca.transform(origEnVecs, origForeignVecs)
'''生成投影后的向量'''
#tmpOutput = rcca._listdot([d.T for d in [origEnVecs, origForeignVecs]], cca.ws)
origEnVecsProjected = preprocessing.normalize(X_c)
#origEnVecsProjected = preprocessing.scale(tmpOutput[0])
origEnVecsProjected = np.column_stack(
(tmp[:, :1], origEnVecsProjected.astype(np.str)))
origForeignVecsProjected = preprocessing.normalize(Y_c)
#origForeignVecsProjected = preprocessing.scale(tmpOutput[1])
origForeignVecsProjected = np.column_stack(
(tmp2[:, :1], origForeignVecsProjected.astype(np.str)))
np.savetxt(outputEnFile, origEnVecsProjected, fmt="%s", delimiter=' ')
np.savetxt(outputForeignFile,
origForeignVecsProjected,
fmt="%s",
delimiter=' ')
print "work over!"
def CCA_project_vectors(args,
src_dico,
tgt_dico,
src_full,
tgt_full,
src_train,
tgt_train,
NUM_dim=100):
print('Exporting embeddings...')
OutputDir = "output/{}-{}/".format(args.src_lang, args.tgt_lang)
if not os.path.exists(OutputDir):
os.makedirs(OutputDir)
cca = CCA(n_components=NUM_dim)
print("Fitting...")
cca.fit(src_train, tgt_train)
print(cca.get_params())
X_c, Y_c = cca.transform(src_full, tgt_full)
src_out, tgt_out = utils.norm_embeddings(X_c), utils.norm_embeddings(Y_c)
print("Exporting embeddings...")
utils.export_embeddings(src_dico[0], src_out,
OutputDir + 'projected.{}'.format(args.src_lang))
utils.export_embeddings(tgt_dico[0], tgt_out,
OutputDir + 'projected.{}'.format(args.tgt_lang))
print("work over!")
#Y = N.array(Z)[:,3:6].tolist() print 'X=\n',X print 'Y=\n',Y Rx = N.corrcoef(X.T) Ry = N.corrcoef(Y.T) cca = CCA(n_components=1) cca.fit(X, Y) print "Rx:\n", Rx print "Ry:\n", Ry print "x_weights:\n", cca.x_weights_ print "y_weights:\n", cca.y_weights_ print "x_loadings:\n", cca.x_loadings_ print "y_loadings:\n", cca.y_loadings_ print "x_scores_:\n", cca.x_scores_ print "y_scores_:\n", cca.y_scores_ loadings_man_x = N.dot(Rx, cca.x_weights_) loadings_man_y = N.dot(Ry, cca.y_weights_) print "loadings_man_x:\n",loadings_man_x print "loadings_man_y:\n",loadings_man_y print cca.get_params() #X_c, Y_c = cca.transform(X, Y) #print "\n\n", X_c,"\n\n", Y_c