Amazon Alexa Review Analysis 2 _ Live - LSTM.py

#!/usr/bin/env python
# coding: utf-8

# In[1]:


import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
get_ipython().run_line_magic('matplotlib', 'inline')
from matplotlib import style
style.use('ggplot')
import re
from nltk.tokenize import word_tokenize
from nltk.stem import PorterStemmer
from nltk.corpus import stopwords
stop_words = set(stopwords.words('english'))
from wordcloud import WordCloud
from sklearn.feature_extraction.text import TfidfVectorizer
from sklearn.model_selection import train_test_split


# In[2]:


df = pd.read_csv('amazon_alexa.tsv', sep='\t')
df.head()


# In[3]:


df.info()


# In[4]:


sns.countplot(x='rating', data=df)


# In[5]:


df['rating'].value_counts()


# In[6]:


fig = plt.figure(figsize=(7,7))
tags = df['rating'].value_counts()
tags.plot(kind='pie', autopct='%1.1f', label='')
plt.title('Distribution of the different ratings')
plt.show()


# In[12]:


fig = plt.figure(figsize=(7,7))
tags = df['rating'].value_counts()
tags.plot(kind='pie', autopct='%1.1f', label='')
plt.title('Distribution of the different ratings')
plt.show()

fig = plt.figure(figsize=(10,10))
colors = ("red", "yellow", "crimson", "blue", "darkgreen") 
wp = {'linewidth':2, 'edgecolor':"black"} 
data = df['rating'].value_counts()
explode = (0.1,0.1,0.1, 0.1,0.1)
classes = ['1','2','3','4','5']
data.plot(kind='pie', autopct='%0.1f%%',labels=classes, startangle = 45, colors = colors,
         explode = explode, shadow=True, wedgeprops = wp)
plt.legend(title="species")
plt.title("Distribution of the different ratings", loc='center', color='black', fontsize='25', fontweight='bold')
plt.show()


# In[7]:


fig = plt.figure(figsize=(20,10))
sns.countplot(y='variation', data=df)


# In[8]:


df['variation'].value_counts()


# In[9]:


sns.countplot(x='feedback', data=df)
plt.show()


# In[10]:


fig = plt.figure(figsize=(7,7))
tags = df['feedback'].value_counts()
tags.plot(kind='pie', autopct="%1.1f%%", label='')
plt.title("Distribution of the different sentiments")
plt.show()


# In[11]:


for i in range(5):
    print(df['verified_reviews'].iloc[i], "\n")
    print(df['feedback'].iloc[i], "\n")


# In[12]:


def data_processing(text):
    text = text.lower()
    text = re.sub(r"https\S+|www\S+|http\S+", '', text, flags = re.MULTILINE)
    text = re.sub(r'[^\w\s]','', text)
    text_tokens = word_tokenize(text)
    filtered_text = [w for w in text_tokens if not w in stop_words]
    return " ".join(filtered_text)


# In[13]:


df.verified_reviews = df['verified_reviews'].apply(data_processing)


# In[14]:


stemmer = PorterStemmer()
def stemming(data):
    text = [stemmer.stem(word) for word in data]
    return data


# In[15]:


df['verified_reviews'] = df['verified_reviews'].apply(lambda x: stemming(x))


# In[16]:


for i in range(5):
    print(df['verified_reviews'].iloc[i], "\n")
    print(df['feedback'].iloc[i], "\n")


# In[17]:


pos_reviews= df[df.feedback==1]
pos_reviews.head()


# In[18]:


text = ' '.join([word for word in pos_reviews['verified_reviews']])
plt.figure(figsize=(20,15), facecolor='None')
wordcloud = WordCloud(max_words=500, width=1600, height=800).generate(text)
plt.imshow(wordcloud, interpolation='bilinear')
plt.axis("off")
plt.title('Most frequent words in positive reviews', fontsize=19)
plt.show()


# In[19]:


neg_reviews= df[df.feedback==0]
neg_reviews.head()


# In[20]:


text = ' '.join([word for word in neg_reviews['verified_reviews']])
plt.figure(figsize=(20,15), facecolor='None')
wordcloud = WordCloud(max_words=500, width=1600, height=800).generate(text)
plt.imshow(wordcloud, interpolation='bilinear')
plt.axis("off")
plt.title('Most frequent words in negative reviews', fontsize=19)
plt.show()


# In[21]:


from keras.models import Sequential
from keras.preprocessing.text import Tokenizer
from keras.preprocessing.sequence import pad_sequences
from keras.layers import Dense, Embedding, LSTM, SpatialDropout1D


# In[22]:


tokenizer = Tokenizer(num_words=500, split=' ')
tokenizer.fit_on_texts(df['verified_reviews'])
X = tokenizer.texts_to_sequences(df['verified_reviews'])
X = pad_sequences(X)
X


# In[24]:


X.shape


# In[26]:


Y = df['feedback']


# In[28]:


x_train, x_test, y_train, y_test = train_test_split(X, Y, test_size=0.2, random_state=42)


# In[29]:


print("Size of x_train: ", (x_train.shape))
print("Size of y_train: ", (y_train.shape))
print("Size of x_test: ", (x_test.shape))
print("Size of y_test: ", (y_test.shape))


# In[30]:


model = Sequential()
model.add(Embedding(500, 120, input_length= X.shape[1]))
model.add(SpatialDropout1D(0.4))
model.add(LSTM(64))
model.add(Dense(1, activation='sigmoid'))
model.compile(loss='binary_crossentropy', optimizer='adam', metrics=['accuracy'])
print(model.summary())


# In[31]:


history = model.fit(x_train, y_train, epochs=10, batch_size=32)


# In[32]:


Scores = model.evaluate(x_test, y_test)
print("Accuracy = %0.3f%%"%(Scores[1]*100))


# In[33]:


plt.figure(figsize=(16,8))
plt.subplot(1,2,1)
plt.plot(history.history['loss'], color='r', label='loss')
plt.title('Training Loss')
plt.xlabel("Number of epochs")
plt.ylabel("Loss")
plt.subplot(1,2,2)
plt.plot(history.history['accuracy'], color='b', label='accuracy')
plt.title('Training accuracy')
plt.xlabel("Number of epochs")
plt.ylabel("Accuracy")
plt.legend()
plt.show()


# ### Model configuration 2 

# In[34]:


model = Sequential()
model.add(Embedding(500, 120, input_length= X.shape[1]))
model.add(LSTM(32))
model.add(Dense(1, activation='sigmoid'))
model.compile(loss='binary_crossentropy', optimizer='adam', metrics=['accuracy'])
print(model.summary())


# In[35]:


history = model.fit(x_train, y_train, epochs=10, batch_size=32)


# In[36]:


Scores = model.evaluate(x_test, y_test)
print("Accuracy = %0.3f%%"%(Scores[1]*100))


# In[37]:


plt.figure(figsize=(16,8))
plt.subplot(1,2,1)
plt.plot(history.history['loss'], color='r', label='loss')
plt.title('Training Loss')
plt.xlabel("Number of epochs")
plt.ylabel("Loss")
plt.subplot(1,2,2)
plt.plot(history.history['accuracy'], color='b', label='accuracy')
plt.title('Training accuracy')
plt.xlabel("Number of epochs")
plt.ylabel("Accuracy")
plt.legend()
plt.show()


# In[ ]: