def bayes_train_test(X,y):
n_splits=3
kf = KFold(n_splits,shuffle=True)
kf.get_n_splits(X)
accuracy=[]
f1_macro=[]
f1_micro=[]
for train_index, test_index in kf.split(X):
X_train, X_test = X[train_index], X[test_index]
y_train, y_test = y[train_index], y[test_index]
y_pred=bayes.predict(X_test,X_train,y_train)
accuracy.append(accuracy_score(y_test,y_pred))
f1_macro.append(f1_score(y_test,y_pred,average='macro'))
f1_micro.append(f1_score(y_test,y_pred,average='micro'))
acc=np.mean(np.asarray(accuracy))
f1_ma=np.mean(np.asarray(f1_macro))
f1_mi=np.mean(np.asarray(f1_micro))
print("Test accuracy ::",acc)
print("Test macro F1 Score ::",f1_ma)
print("Test micro F1 Score ::",f1_mi)
return acc,f1_ma,f1_mi
def classify_bayes():
fracs = [x / 10 for x in range(1, 11)]
digitd = (read_digitdata('training'), read_digitdata('test'),
read_digitdata('validation'))
faced = (read_facedata('train'), read_facedata('test'),
read_facedata('validation'))
datasets = [faced, digitd]
# Bayesian
for dataset in datasets:
for f in fracs:
fs = take_sample(dataset[0], f)
t1 = time.time()
model = bayes.train(fs)
t2 = time.time()
total = len(dataset[1])
cor = 0
for item in dataset[1]:
cl, logprob = bayes.predict(item[0], model)
if cl == item[1]:
cor += 1
acc = float(cor) / total
print("Bayes class accuracy frac {} train_time {} accuracy {}".
format(f, t2 - t1, acc))
test["Saltratio"]=0.0
test["Saltratio"].ix[test["gender"]=="male"]=test["Salt[g]"]/8.0
test["Saltratio"].ix[test["gender"]=="female"]=test["Salt[g]"]/7.0
test=test.drop(["gender","age","height","weight","EER[kcal]",
"P target(15%)[g]","F target(25%)[g]","C target(60%)[g]",
"E[kcal]","P[g]","F[g]","C[g]","Salt[g]"],axis=1)
#---transform dataframe to list-----------------------------------------
test_array=np.array(test)
test_list=test_array.tolist()
#---predict the class and get a list of classes------------------------
predictions = []
for i in range(len(test_list)):
if test_list[i][0]=="breakfast":
del test_list[i][0]
prediction = bayes.predict(train_breakfast_dict,
train_breakfast_summaries,test_list[i])
if test_list[i][0]=="lunch":
del test_list[i][0]
prediction = bayes.predict(train_lunch_dict,
train_lunch_summaries,test_list[i])
if test_list[i][0]=="dinner":
del test_list[i][0]
prediction = bayes.predict(train_dinner_dict,
train_dinner_summaries,test_list[i])
predictions.append(prediction)
#---write the claseees into the table ------------------------------------
score = pd.DataFrame(predictions,columns=["scores"])
test = pd.read_excel('filename of test data')
test=test.reset_index(drop=True)
testscore = pd.concat([test,score],axis=1)
testscore.to_excel('filename of the test data with results')
import bayes as by
import numpy as ny
test_doc1=['love','my','dalmation']
test_doc2=['love','dog','my','dog']
dataset,labels=by.parseFile('data')
vocab_list=by.creatVocabList(dataset)
test_doc1=ny.array(by.setWords2Bag2(vocab_list, test_doc2))
pVec,pDoc,index=by.trainNB(dataset, labels,vocab_list)
result=by.predict(pVec,pDoc,index,test_doc1)
print(result)
"gender", "age", "height", "weight", "EER[kcal]", "Type",
"P target(15%)[g]", "F target(25%)[g]", "C target(60%)[g]", "E[kcal]",
"P[g]", "F[g]", "C[g]", "Salt[g]"
],
axis=1)
accuracy = 0
for iter_num in range(1000):
#---Validation Split-----------------------------------------------------------
train, val = train_test_split(trainval, test_size=0.1)
val_x = val.drop(["Score(1:worst 2:bad 3:good 4:best)"], axis=1)
val_x_array = np.array(val_x)
val_x_list = val_x_array.tolist()
#---devide train data according to class-------------------------------
train = np.array(train)
train = train.tolist()
train_dict = bayes.fea_and_class(train)
#---calculate the mean and std of each feature according to category-------
train_summaries = bayes.summarizeByClass(train_dict)
#---predict the class and get a list of classes----------------------
predictions = []
for i in range(len(val_x_list)):
prediction = bayes.predict(train_dict, train_summaries, val_x_list[i])
predictions.append(prediction)
#calculate the accuracy using the validation and predictions---------------
val_array = np.array(val)
val_list = val_array.tolist()
accuracy += bayes.accuracy(val_list, predictions)
print accuracy / 1000.0