def predict(language_dist,comment):
nLanguages=3
n=3
LANGUAGES=list(set(['fr','es','ru']))
print("implementing SVM ranking")
X=np.loadtxt('X.txt')
Y=np.loadtxt('Y.txt')
weights=SVM(X,Y,nLanguages,n)
developmentData=loader.loaderShuffleDataTest()
computeFeatures(developmentData,language_dist,LANGUAGES,1)
Ypredict=weights.predict(comment)
if Ypredict==0:
return "French"
if Ypredict==1:
return "Spanish"
else:
return "Russian"
precision=nCorrect/len(Y)
print("Precision is equal to "+str(precision)+"on "+str(nbPrediction)+" comments")
print(str(ncorrectEnglish)+" english correct on "+str(nEnglish))
print(str(ncorrectFrench)+" english correct on "+str(nFrench))
print(str(ncorrectRussian)+" english correct on "+str(nRussian))
# for i in range(0,n):
# j=0
# for language in LANGUAGES:
# NumbersWords[i,j]=sum(language_dist[language]['words'][i+1].values())
# j=j+1
# #print NumbersWords
# #training features
# print ("Starting loaderShuffleData")
# trainData=loader.loaderShuffleData()
# print ("computing X and Y matrices")
# computeFeatures(trainData,language_dist,LANGUAGES,NumbersWords,0)
print("implementing SVM ranking")
X=np.loadtxt('Xtrain.txt')
Y=np.loadtxt('Ytrain.txt')
weights=SVM(X,Y,nLanguages,n)
developmentData=loader.loaderShuffleDataTest()
computeFeatures(developmentData,language_dist,LANGUAGES,1)
X=np.loadtxt('Xtest.txt')
Y=np.loadtxt('Ytest.txt')
nbPrediction=len(Y)
Ypredict=weights.predict(X)
precision=0
nCorrect=0
for i in range(len(Y)):
if Ypredict[i]==Y[i]:
nCorrect=nCorrect+1
precision=nCorrect/len(Y)
print("Precision is equal to "+str(precision)+"on "+str(nbPrediction)+" comments")
# LANGUAGES = list(set(['english', 'french', 'german']))
# for language in LANGUAGES: