12/2/21, 9:51 AM sentimen2.

ipynb - Colaboratory

1 from google.colab import drive
2 drive.mount('/content/drive')

Mounted at /content/drive

1 import pandas as pd
2 import numpy as np
3 import re
4 import nltk
5 from nltk.corpus import stopwords
6
7
8
9 from numpy import array
10 from keras.preprocessing.text import one_hot
11 from keras.preprocessing.sequence import pad_sequences
12 from keras.models import Sequential
13 from keras.layers.core import  Activation, Dropout, Dense
14 from keras.layers import  Flatten
15 from keras.layers import GlobalMaxPooling1D
16
17 from keras.layers.embeddings import Embedding
18 from sklearn.model_selection import train_test_split
19 from keras.preprocessing.text import Tokenizer

1 #input data
2 data = pd.read_excel('/content/drive/MyDrive/sentiment.xlsx')
3 data.isnull().values.any()
4 data.shape
5

(1005, 2)

1
#Menampilkan Data 5 baris

2
data.head()

content sentiment

0 Saya baru kemaren instal aplikasi ini langsung... Negatif

1 Baru diupdate mlah ngelag terus pdahal jaringa... Negatif

2 Kecewa dengan aplikasinya. Disaat saya ingin b... Negatif

3 Kiye apik yakin Positif

4 Bapak saya mau pesen 3 penumpang, saat pemilih... Negatif

https://colab.research.google.com/github/BudiAtu/BudiAtu.github.io/blob/main/sentimen2.ipynb#printMode=true 1/12
12/2/21, 9:51 AM sentimen2.ipynb - Colaboratory

1
#contoh menampilkan data attribut dari record 3

2
data["content"][3]

'Kiye apik yakin'

1
#Seaborn merupakan pustaka visualisasi data 

2
import seaborn as sns

3
sns.countplot(x='sentiment', data=data)

<matplotlib.axes._subplots.AxesSubplot at 0x7f6880c4a210>

1
#preprosesing data

2
def preprocess_text(sen):

3
  #Removing html tags

4
  sentence = remove_tags(sen)

6
  #removing punctuation and numbers

7
  sentence = re.sub('[^a-zA-Z]', ' ', sentence)

9
  #single character removal

10
  sentence = re.sub(r"\s+[a-zA-Z]\s+", ' ', sentence)

11

12

13

14
  #Removing mutiple spaces

15
  sentence = re.sub(r'\s+', ' ', sentence)

16

17
  return sentence

1
#eretan karakter yang digunakan untuk pencarian string atau teks dengan menggunakan pola (
2
import re

3
TAG_RE = re.compile(r'<[^>]+>')

4
def remove_tags(text):

5
  return TAG_RE.sub('', text)

1
#membuat array list

https://colab.research.google.com/github/BudiAtu/BudiAtu.github.io/blob/main/sentimen2.ipynb#printMode=true 2/12
12/2/21, 9:51 AM sentimen2.ipynb - Colaboratory

2
X = []

3
sentences = list(data['content'])

4
for sen in sentences:

5
  X.append(preprocess_text(sen))

1
X[0]

'Saya baru kemaren instal aplikasi ini langsung Bisa dibuka Tapi hari ini buka lagi ko g
ncul layar putih Ditunggu lama banget sampe saya uinstall lalu instal lagi tetep aja kek
iperbaiki dong Ga malah memudahkan malah ribet jadinya Kita komplain juga ga malah dibal

1
#membuat array dari sentimen data

2
y = data['sentiment']

3
y = np.array(list(map(lambda x: 1 if x=="positif" else 0, y)))

1
#split data untuk membagi data training dan data testing

2
from sklearn.model_selection import train_test_split

3
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.20, random_state=42)

4
print('Data Train : ',X_train)

5
print('Data Test : ',X_test)

Data Train : ['Sangat membantu dalam pembelian tiket kereta api tanpa harus antri di lo
Data Test : ['mau daftar susah ga dapet verifikasi ', 'Kak ini gimna pesen tiket udah d

1
#memisahkan teks menjadi potongan-potongan berupa token, bisa berupa potongan huruf, kata, 
2
tokenizer = Tokenizer(num_words=5000)

3
tokenizer.fit_on_texts(X_train)

5
X_train = tokenizer.texts_to_sequences(X_train)

6
X_test = tokenizer.texts_to_sequences(X_test)

1
#adding 1 because of reserved 0 index

2
vocab_size = len(tokenizer.word_index) + 1

3
maxlen = 100

4
X_train = pad_sequences(X_train, padding='post', maxlen=maxlen)
5
X_test = pad_sequences(X_test, padding='post', maxlen=maxlen)

1
from numpy import array

2
from numpy import  asarray

3
from numpy import zeros

5
embeddings_dictionary = dict()
6
glove_file =open('/content/drive/MyDrive/glove.6B.100d.txt', encoding="utf8")

7
for line in glove_file:

8
  records =line.split()

https://colab.research.google.com/github/BudiAtu/BudiAtu.github.io/blob/main/sentimen2.ipynb#printMode=true 3/12
12/2/21, 9:51 AM sentimen2.ipynb - Colaboratory

9
  word = records[0]

10
  vector_dimensions = asarray(records[1:], dtype='float32')

11
  embeddings_dictionary[word] = vector_dimensions

12
glove_file.close()

1
embedding_matrix = zeros((vocab_size, 100))

2
for word, index in tokenizer.word_index.items():

3
  embedding_vector = embeddings_dictionary.get(word)

4
  if embedding_vector is not None:

5
    embedding_matrix[index] = embedding_vector

1
#Dekrasi untuk Model CNN

2
from keras.models import Model

3
from keras.layers import Conv1D, Dense, MaxPool1D, Flatten, Input

4
from tensorflow.keras.optimizers import Adam

5
from tensorflow.keras.utils import plot_model

7
import numpy as np

8
model = Sequential()

1
#membangun model CNN

2
embedding_layer = Embedding(vocab_size, 100, weights=[embedding_matrix], input_length=maxl
3
model.add(embedding_layer)

4
model.add(Conv1D(128, 5, activation='relu'))

5
model.add(GlobalMaxPooling1D())

6
model.add(Dense(1, activation='sigmoid'))

7
model.compile(optimizer='adam', loss='binary_crossentropy', metrics=['acc'])

1
import os

2
os.mkdir("/content/simpan_model")

1
#Menampilkan hasil pembuatan MOdel

2
model.summary()       

3
plot_model(model, to_file='model.png', show_shapes=True, show_layer_names=False, rankdir='

https://colab.research.google.com/github/BudiAtu/BudiAtu.github.io/blob/main/sentimen2.ipynb#printMode=true 4/12
12/2/21, 9:51 AM sentimen2.ipynb - Colaboratory

Model: "sequential_18"

_________________________________________________________________

Layer (type) Output Shape Param #

=================================================================

embedding_25 (Embedding) (None, 100, 100) 231400

conv1d_25 (Conv1D) (None, 96, 128) 64128

global_max_pooling1d_18 (Gl (None, 128) 0

obalMaxPooling1D)

dense_18 (Dense) (None, 1) 129

=================================================================

Total params: 295,657

Trainable params: 64,257

Non-trainable params: 231,400

_________________________________________________________________

1
#melakukan Training

2
import pandas as pd

3
history = model.fit(X_train, y_train, validation_data=(X_test,y_test), batch_size=128, epo
4
model.save("mymodel.h5")

Epoch 1/10

6/6 [==============================] - 1s 128ms/step - loss: 0.3825 - acc: 0.8686 - val_

6/6 [==============================] - 1s 93ms/step - loss: 0.0741 - acc: 0.9986 - val_l

6/6 [==============================] - 1s 95ms/step - loss: 0.0397 - acc: 0.9986 - val_l

6/6 [==============================] - 1s 88ms/step - loss: 0.0296 - acc: 1.0000 - val_l

6/6 [==============================] - 1s 93ms/step - loss: 0.0250 - acc: 1.0000 - val_l

6/6 [==============================] - 1s 90ms/step - loss: 0.0225 - acc: 1.0000 - val_l

6/6 [==============================] - 1s 94ms/step - loss: 0.0209 - acc: 1.0000 - val_l

6/6 [==============================] - 1s 91ms/step - loss: 0.0198 - acc: 1.0000 - val_l

6/6 [==============================] - 1s 92ms/step - loss: 0.0188 - acc: 1.0000 - val_l

https://colab.research.google.com/github/BudiAtu/BudiAtu.github.io/blob/main/sentimen2.ipynb#printMode=true 5/12
12/2/21, 9:51 AM sentimen2.ipynb - Colaboratory

Epoch 10/10

6/6 [==============================] - 1s 90ms/step - loss: 0.0181 - acc: 1.0000 - val_l

1
#testing model

2
score = model.evaluate(X_train, y_train)

3
print('Test Accuracy: %f' % (acc*100))

26/26 [==============================] - 0s 10ms/step - loss: 0.0163 - acc: 1.0000

Test Accuracy: 100.000000

1
# evaluate

2
score = model.evaluate(X_test, y_test, verbose=1)

3
print('Test Accuracy: %f' % (acc*100))

7/7 [==============================] - 0s 11ms/step - loss: 0.0202 - acc: 1.0000

Test Accuracy: 100.000000

1
# Evaluating the model's accuracy

2
cnn_score = model.evaluate(X_test, y_test, verbose=1)

3
print("\nTEST SCORES OF MODEL")

4
print("Test loss:", cnn_score[0])

5
print("Test Accuracy:", cnn_score[1])

7/7 [==============================] - 0s 10ms/step - loss: 0.0202 - acc: 1.0000

TEST SCORES OF MODEL

Test loss: 0.02021356299519539

Test Accuracy: 1.0

1
# Final Results

2
print("Convolution Neural Network")

3
print("Traning Accuracy : {}" .format(history.history['acc']))

4
print("Validation Accuracy : {}" .format(history.history['val_acc']))

5
print("\nTraning Loss : {}" .format(history.history['loss']))

6
print("Validation Loss : {}" .format(history.history['val_loss']))

Convolution Neural Network

Traning Accuracy : [0.8686030507087708, 0.9986168742179871, 0.9986168742179871, 1.0, 1.0

Traning Loss : [0.3824833929538727, 0.07414576411247253, 0.03972981497645378, 0.02959362

1
#visualisasi data hasil 

3
import matplotlib.pyplot as plt

4
plt.plot(history.history['acc'])

https://colab.research.google.com/github/BudiAtu/BudiAtu.github.io/blob/main/sentimen2.ipynb#printMode=true 6/12
12/2/21, 9:51 AM sentimen2.ipynb - Colaboratory

6
plt.plot(history.history['val_acc'])

8
plt.title('model accuracy')

9
plt.ylabel('accuracy')

10
plt.xlabel('epoch')

11
plt.legend(['train', 'test'], loc='upper left')

12
plt.show()

13

14
plt.plot(history.history['loss'])

15

16
plt.plot(history.history['val_loss'])

17

18
plt.title('model loss')

19
plt.ylabel('loss')

20
plt.xlabel('epoch')

21
plt.legend(['train', 'test'], loc='upper left')

22
plt.show()

1
y_pred=model.predict(X_test)

https://colab.research.google.com/github/BudiAtu/BudiAtu.github.io/blob/main/sentimen2.ipynb#printMode=true 7/12
12/2/21, 9:51 AM sentimen2.ipynb - Colaboratory

2
y_pred

array([[1.0196775e-02],

[1.1258881e-04],

[3.4761429e-04],

[7.8367775e-05],

[6.1708874e-05],

[2.1058321e-04],

[2.8647717e-05],

[2.8658837e-02],

[4.4444203e-04],

[2.8818846e-04],

[2.9569298e-02],

[4.2870186e-05],

[3.8734078e-04],

[6.0484995e-06],

[6.9963932e-03],

[3.5810471e-03],

[7.5332224e-03],

[5.1296443e-02],

[2.4898305e-05],

[1.4734268e-04],

[4.5093942e-01],

[3.1632185e-04],

[9.2998516e-06],

[1.6399613e-06],

[1.4100856e-01],

[6.1747432e-04],

[3.7493706e-03],

[4.9095255e-05],

[6.7559566e-05],

[1.5625358e-04],

[2.2199750e-04],

[1.0730326e-03],

[1.7580940e-05],

[8.2913339e-03],

[4.5093942e-01],

[1.9695372e-05],

[1.0849029e-02],

[4.0538736e-05],

[5.5712094e-06],

[1.4087558e-04],

[2.2138178e-02],

[9.9324882e-03],

[8.4398496e-05],

[7.3028594e-02],

[3.8266182e-04],

[1.6280830e-05],

[3.2701031e-05],

[1.3604760e-04],

[3.7491322e-04],

[6.3771009e-04],

[1.4735192e-02],

[1.5205215e-05],

[2.3714689e-05],

[3.7834048e-04],

https://colab.research.google.com/github/BudiAtu/BudiAtu.github.io/blob/main/sentimen2.ipynb#printMode=true 8/12
12/2/21, 9:51 AM sentimen2.ipynb - Colaboratory

[7.9051220e-05],

[2.8041005e-03],

[5.7262852e-05],

[2.0313615e-05],

1
y_pred=np.argmax(X_test,axis=1)

2
y_pred

array([ 4, 2, 2, 12, 16, 10, 8, 1, 5, 37, 3, 10, 0, 0, 5, 1, 0,

3, 10, 5, 0, 3, 20, 17, 2, 6, 2, 30, 9, 0, 0, 4, 17, 12,

1, 14, 2, 19, 10, 16, 2, 1, 12, 2, 9, 16, 5, 6, 7, 3, 7,

5, 9, 2, 9, 3, 17, 34, 2, 31, 1, 5, 10, 0, 5, 0, 11, 7,

5, 1, 14, 13, 14, 2, 0, 10, 2, 0, 14, 10, 7, 15, 30, 6, 6,

0, 7, 1, 0, 2, 2, 1, 1, 11, 0, 5, 7, 4, 14, 28, 2, 4,

6, 6, 3, 10, 3, 3, 4, 1, 2, 4, 13, 3, 8, 15, 6, 6, 14,

6, 22, 4, 1, 3, 26, 1, 20, 3, 11, 4, 6, 5, 2, 0, 59, 12,

23, 3, 1, 14, 9, 14, 15, 77, 5, 0, 1, 11, 2, 14, 6, 4, 5,

8, 14, 6, 1, 3, 10, 34, 12, 5, 7, 13, 11, 6, 8, 2, 4, 9,

21, 16, 16, 20, 35, 11, 3, 17, 7, 4, 7, 1, 12, 1, 13, 4, 8,

5, 10, 2, 0, 2, 6, 4, 3, 5, 18, 0, 18, 15, 2])

1
from sklearn.metrics import confusion_matrix,classification_report

2
import seaborn as sns

1
cm=confusion_matrix(y_test,y_pred)

2
print("Accuracy : {:.3f}%".format(acc*100))

3
print("Loss     : {:.3f}".format(loss))

5
print("\nClassification Report :\n")

6
ax=sns.heatmap(cm,annot=True,cmap='Blues',fmt=' ')

7
ax.set_title('Confusion Matrix')

8
ax.set_xlabel('y_test')

9
ax.set_ylabel('y_pred')

https://colab.research.google.com/github/BudiAtu/BudiAtu.github.io/blob/main/sentimen2.ipynb#printMode=true 9/12
12/2/21, 9:51 AM sentimen2.ipynb - Colaboratory

Accuracy : 100.000%

Loss : 0.010

Classification Report :

1
print(classification_report(y_test,y_pred))

Text(33.0, 0.5, 'y_pred')

0 1.00 0.08 0.16 201

1 0.00 0.00 0.00 0

2 0.00 0.00 0.00 0

3 0.00 0.00 0.00 0

4 0.00 0.00 0.00 0

5 0.00 0.00 0.00 0

6 0.00 0.00 0.00 0

7 0.00 0.00 0.00 0

8 0.00 0.00 0.00 0

9 0.00 0.00 0.00 0

10 0.00 0.00 0.00 0

11 0.00 0.00 0.00 0

12 0.00 0.00 0.00 0

13 0.00 0.00 0.00 0

14 0.00 0.00 0.00 0

15 0.00 0.00 0.00 0

16 0.00 0.00 0.00 0

17 0.00 0.00 0.00 0

18 0.00 0.00 0.00 0

19 0.00 0.00 0.00 0

20 0.00 0.00 0.00 0

21 0.00 0.00 0.00 0

22 0.00 0.00 0.00 0

23 0.00 0.00 0.00 0

26 0.00 0.00 0.00 0

28 0.00 0.00 0.00 0

30 0.00 0.00 0.00 0

31 0.00 0.00 0.00 0

34 0.00 0.00 0.00 0

35 0.00 0.00 0.00 0

37 0.00 0.00 0.00 0

59 0.00 0.00 0.00 0

77 0.00 0.00 0.00 0

accuracy 0.08 201

macro avg 0.03 0.00 0.00 201

weighted avg 1.00 0.08 0.16 201

/usr/local/lib/python3.7/dist-packages/sklearn/metrics/_classification.py:1308: Undefine
_warn_prf(average, modifier, msg_start, len(result))

/usr/local/lib/python3.7/dist-packages/sklearn/metrics/_classification.py:1308: Undefine
_warn_prf(average, modifier, msg_start, len(result))

/usr/local/lib/python3.7/dist-packages/sklearn/metrics/_classification.py:1308: Undefine
_warn_prf(average, modifier, msg_start, len(result))

https://colab.research.google.com/github/BudiAtu/BudiAtu.github.io/blob/main/sentimen2.ipynb#printMode=true 10/12
12/2/21, 9:51 AM sentimen2.ipynb - Colaboratory

1
# Getting input from user

2
print("\nMAKING PREDICTION ON SINGLE INSTANCE USING THREE MODELS THAT WE HAVE BUILT")

3
reviewNo = int(input("\nEnter the review number whose sentiment is to be predcited"))

4
reviewText = X[reviewNo]

5
print("\nCONTENT OF THE REVIEW CHOSEN IS")

6
print(reviewText)

8
# Converting text to numeric form

9
    #Using the tokenizer built earlier

10
#Since we have trained with a list of reviews and now we are feeding in a string, we need 
11
reviewProcessed = tokenizer.texts_to_sequences(reviewText)

12
print("\nINTERGER SEQUENCE OF THE REVIEW CHOSEN IS")

13
print(reviewProcessed)

14

15
flat_list = []

16
for sublist in reviewProcessed:

17
    for item in sublist:

18
        flat_list.append(item)
19
#Making the entire items as a single list

20
flat_list = [flat_list]

21
#Padding to fit into model

22
reviewSequence = pad_sequences(flat_list, padding='post', maxlen=maxlen)

MAKING PREDICTION ON SINGLE INSTANCE USING THREE MODELS THAT WE HAVE BUILT

Enter the review number whose sentiment is to be predcited7

CONTENT OF THE REVIEW CHOSEN IS

Udah punya akun mau login kalimatnya aduh timeout Lah ini kai access atau lagu dangdut K

INTERGER SEQUENCE OF THE REVIEW CHOSEN IS

[[], [2201], [], [], [], [1424], [], [], [], [], [], [], [], [], [], [], [], [], [], [],

https://colab.research.google.com/github/BudiAtu/BudiAtu.github.io/blob/main/sentimen2.ipynb#printMode=true 11/12
12/2/21, 9:51 AM sentimen2.ipynb - Colaboratory

https://colab.research.google.com/github/BudiAtu/BudiAtu.github.io/blob/main/sentimen2.ipynb#printMode=true 12/12