import necessary library¶

In [1]:
import numpy as np
import pandas as pd
pd.set_option('display.max_columns', None)
import matplotlib.pyplot as plt
import seaborn as sns

load dataset¶

In [2]:
df = pd.read_csv('Restaurant-Reviews.tsv', sep='\t', quoting=3)

exploratory data analysis¶

In [3]:
df.head()
Out[3]:
Review Liked
0 Wow... Loved this place. 1
1 Crust is not good. 0
2 Not tasty and the texture was just nasty. 0
3 Stopped by during the late May bank holiday of... 1
4 The selection on the menu was great and so wer... 1
In [4]:
df.tail()
Out[4]:
Review Liked
995 I think food should have flavor and texture an... 0
996 Appetite instantly gone. 0
997 Overall I was not impressed and would not go b... 0
998 The whole experience was underwhelming, and I ... 0
999 Then, as if I hadn't wasted enough of my life ... 0
In [5]:
df.isna().sum()
Out[5]:
Review    0
Liked     0
dtype: int64
In [6]:
print(len(df))
print(len(df[df['Liked'] == 1]))

1000
500

data cleaning¶

In [7]:
import re
import nltk
In [8]:
nltk.download('stopwords')

[nltk_data] Downloading package stopwords to
[nltk_data]     C:\Users\yandiher\AppData\Roaming\nltk_data...
[nltk_data]   Package stopwords is already up-to-date!
Out[8]:
True
In [9]:
from nltk.corpus import stopwords
In [10]:
df['Review'][0]
Out[10]:
'Wow... Loved this place.'
In [11]:
review = re.sub(pattern='[^a-zA-Z]',
                repl=' ',
                string=df['Review'][0]).lower().split()
In [12]:
preview = []
for word in review:
    if word not in stopwords.words('english'):
        preview.append(word)
print(preview)
print(review)

['wow', 'loved', 'place']
['wow', 'loved', 'this', 'place']
In [13]:
review = [word for word in review if word not in stopwords.words(fileids='english')]
In [14]:
from nltk.stem import PorterStemmer
pStemmer = PorterStemmer()
In [15]:
review = [pStemmer.stem(word=word) for word in review]
In [16]:
review = ' '.join(review)
review
Out[16]:
'wow love place'

bags of word model¶

In [17]:
corpus = []
for i in range (len(df)):
    review = re.sub(pattern='[^a-zA-Z]',
                    repl=' ',
                    string=df['Review'][i])
    review = review.lower()
    review = review.split()
    review = [pStemmer.stem(word) for word in review if word not in stopwords.words(fileids='english')]
    review = ' '.join(review)

    corpus.append(review)
In [18]:
print(corpus[6])
print(df['Review'].iloc[6])

honeslti tast fresh
Honeslty it didn't taste THAT fresh.)
In [19]:
from sklearn.feature_extraction.text import CountVectorizer
vectorizer = CountVectorizer()
vectorizer.fit(corpus)
Out[19]:
CountVectorizer()
In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook.
On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.
CountVectorizer()
In [20]:
x = vectorizer.transform(corpus).toarray()
y = df['Liked'].values

split data¶

In [21]:
from sklearn.model_selection import train_test_split
In [22]:
xTrain, xTest, yTrain, yTest = train_test_split(x, y, train_size=0.8, random_state=42, stratify=y)

build algorithm¶

In [23]:
from sklearn.neural_network import MLPClassifier
model = MLPClassifier()
model.fit(xTrain, yTrain)
prediction = model.predict(xTest)

evaluation¶

In [24]:
from sklearn.metrics import accuracy_score, classification_report, ConfusionMatrixDisplay
In [25]:
print('the accuracy of the model is ', accuracy_score(y_true=yTest, y_pred=prediction))
print(classification_report(y_true=yTest, y_pred=prediction))

the accuracy of the model is  0.77
              precision    recall  f1-score   support

           0       0.79      0.73      0.76       100
           1       0.75      0.81      0.78       100

    accuracy                           0.77       200
   macro avg       0.77      0.77      0.77       200
weighted avg       0.77      0.77      0.77       200
In [26]:
ConfusionMatrixDisplay.from_predictions(y_true=yTest, y_pred=prediction)
plt.show()
In [27]:
def pred(text):
    text = vectorizer.transform(text).toarray()
    text = model.predict(text)
    if test == 1:
        print('the customer(s) love our restaurant.')
    else:
        print('the customer(s) do not like our restaurant.')
In [28]:
test = ['how is it possible that there was a cockroach inside my meal?']
In [29]:
pred(test)

the customer(s) do not like our restaurant.