This project demonstrates text classification using Natural Language Processing (NLP) on restaurant reviews. It involves preprocessing text data, building a Bag of Words model, and applying various machine learning classifiers, including Naive Bayes, SVM, Random Forest, Logistic Regression, and K-Nearest Neighbors. The project also features a 3D visualization of the confusion matrix to analyze model performance.

• Installation
• Usage
• Results
• Visualization

To get started with this project, ensure you have Python installed. Then, install the required libraries using:

pip install numpy pandas matplotlib scikit-learn nltk

1. Import Libraries

   Import the necessary libraries:

   import numpy as np
   import matplotlib.pyplot as plt
   import pandas as pd
   import re
   import nltk
   from nltk.corpus import stopwords
   from nltk.stem.porter import PorterStemmer
   from sklearn.feature_extraction.text import CountVectorizer
   from sklearn.model_selection import train_test_split
   from sklearn.metrics import confusion_matrix, accuracy_score
   from sklearn.naive_bayes import GaussianNB
   from sklearn.svm import SVC
   from sklearn.ensemble import RandomForestClassifier
   from sklearn.linear_model import LogisticRegression
   from sklearn.neighbors import KNeighborsClassifier

2. Import the Dataset

   Load the dataset:

   dataset = pd.read_csv('Restaurant_Reviews.tsv', delimiter='\t', quoting=3)

3. Clean the Texts

   Process the reviews:

   nltk.download('stopwords')
   corpus = []
   for i in range(0, 1000):
       review = re.sub('[^a-zA-Z]', ' ', dataset['Review'][i])
       review = review.lower()
       review = review.split()
       ps = PorterStemmer()
       all_stopwords = stopwords.words('english')
       all_stopwords.remove('not')
       review = [ps.stem(word) for word in review if not word in set(all_stopwords)]
       review = ' '.join(review)
       corpus.append(review)

4. Create the Bag of Words Model

   Convert text to a format suitable for ML models:

   cv = CountVectorizer(max_features=1500)
   X = cv.fit_transform(corpus).toarray()
   y = dataset.iloc[:, -1].values

5. Split the Dataset

   Divide the data into training and test sets:

   X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.20, random_state=0)

6. Train and Evaluate Classifiers

   Train and evaluate various classifiers:

   from sklearn.metrics import confusion_matrix, accuracy_score
   
   # Import Naive Bayes classifier.
   from sklearn.naive_bayes import GaussianNB
   classifierNB = GaussianNB()
   classifierNB.fit(X_train, y_train)
   weightNB = accuracy_score(y_test, classifierNB.predict(X_test))
   
   # Import Support Vector Classifier.
   from sklearn.svm import SVC
   classifierSVM = SVC(kernel='linear', random_state=0)  # Initialize and train SVC with a linear kernel.
   classifierSVM.fit(X_train, y_train)
   weightSVM = accuracy_score(y_test, classifierSVM.predict(X_test))
   
   # Import Random Forest classifier.
   from sklearn.ensemble import RandomForestClassifier
   classifierRF = RandomForestClassifier(n_estimators=1000, criterion='entropy', random_state=42)
   classifierRF.fit(X_train, y_train)
   weightRF = accuracy_score(y_test, classifierRF.predict(X_test))
   
   # Import Logistic Regression classifier.
   from sklearn.linear_model import LogisticRegression
   classifierLR = LogisticRegression(random_state=0)
   classifierLR.fit(X_train, y_train)
   weightLR = accuracy_score(y_test, classifierLR.predict(X_test))
   
   # Import K-Nearest Neighbors classifier.
   from sklearn.neighbors import KNeighborsClassifier
   classifierKNN = KNeighborsClassifier(n_neighbors=5, metric='minkowski', p=2)
   classifierKNN.fit(X_train, y_train)
   weightKNN = accuracy_score(y_test, classifierKNN.predict(X_test))

7. Combine Predictions

   Aggregate predictions from all classifiers:

   # Combine predictions from all classifiers using weighted voting
   weightAll = weightKNN + weightLR + weightNB + weightRF + weightSVM  # Sum of weights of all classifiers
   threshold = 0.4  # Threshold for deciding the final prediction
   y_pred = 1 * (weightNB * classifierNB.predict(X_test) +  # Aggregate predictions with weights
              weightRF * classifierRF.predict(X_test) +
              weightLR * classifierLR.predict(X_test) +
              weightKNN * classifierKNN.predict(X_test) +
              weightSVM * classifierSVM.predict(X_test)) > threshold * weightAll  # Apply threshold

cm = confusion_matrix(y_test, y_pred) # Compute confusion matrix print(cm) # Print confusion matrix accuracy_score(y_test, y_pred) # Print accuracy score

weightAll = weightKNN + weightLR + weightNB + weightRF + weightSVM
threshold = 0.4
y_pred = 1 * (weightNB * classifierNB.predict(X_test) +
              weightRF * classifierRF.predict(X_test) +
              weightLR * classifierLR.predict(X_test) +
              weightKNN * classifierKNN.predict(X_test) +
              weightSVM * classifierSVM.predict(X_test)) > threshold * weightAll

Evaluate the combined model's performance:

cm = confusion_matrix(y_test, y_pred)
print(cm)
print(accuracy_score(y_test, y_pred))

The ensemble model achieved an overall accuracy of 81.5%. This improved performance highlights the effectiveness of combining multiple classifiers—Naive Bayes, SVM, Random Forest, Logistic Regression, and K-Nearest Neighbors—into a single ensemble model. By leveraging the strengths of each classifier, the ensemble approach enhances the accuracy and robustness of predictions compared to using individual classifiers alone.

Visualize the confusion matrix in 3D:

from mpl_toolkits.mplot3d import Axes3D

fig = plt.figure(figsize=(10, 7))
ax = fig.add_subplot(111, projection='3d')

xpos, ypos = np.meshgrid(np.arange(cm.shape[0]), np.arange(cm.shape[1]), indexing="ij")
xpos = xpos.ravel()
ypos = ypos.ravel()
zpos = np.zeros_like(xpos)

dx = dy = 0.5
dz = cm.ravel()

colors = plt.cm.viridis(0.45*dz / np.max(dz))

ax.bar3d(xpos, ypos, zpos, dx, dy, dz, zsort='average', color=colors, edgecolor='black')
ax.set_xlabel('Actual Label')
ax.set_ylabel('Predicted Label')
ax.set_zlabel('Count')
ax.set_xticks(np.arange(cm.shape[0]) + dx / 2)
ax.set_xticklabels(['Negative', 'Positive'])
ax.set_yticks(np.arange(cm.shape[1]) + dy / 2)
ax.set_yticklabels(['Negative', 'Positive'])
plt.title('3D Visualization of Confusion Matrix')
ax.view_init(elev=20, azim=130)
plt.show()

Below is the graphical representation of the confusion matrix.

A 3D visualization of the confusion matrix is provided to illustrate the model's performance