Sentiment Analysis

Overview

Sentiment analysis determines emotional tone and opinions in text, enabling understanding of customer satisfaction, brand perception, and feedback analysis.

Approaches

• Lexicon-based: Using sentiment dictionaries
• Machine Learning: Training classifiers on labeled data
• Deep Learning: Neural networks for complex patterns
• Aspect-based: Sentiment about specific features
• Multilingual: Non-English text analysis

Sentiment Types

• Positive: Favorable, satisfied
• Negative: Unfavorable, dissatisfied
• Neutral: Factual, no clear sentiment
• Mixed: Combination of sentiments

Implementation with Python

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns
from sklearn.feature_extraction.text import TfidfVectorizer
from sklearn.naive_bayes import MultinomialNB
from sklearn.pipeline import Pipeline
from sklearn.model_selection import train_test_split
from sklearn.metrics import classification_report, confusion_matrix, accuracy_score
import re
from collections import Counter

# Sample review data
reviews_data = [
    "This product is amazing! I love it so much.",
    "Terrible quality, very disappointed.",
    "It's okay, nothing special.",
    "Best purchase ever! Highly recommend.",
    "Worst product I've ever bought.",
    "Pretty good, satisfied with the purchase.",
    "Excellent service and fast delivery.",
    "Poor quality and bad customer support.",
    "Not bad, does what it's supposed to.",
    "Absolutely fantastic! Five stars!",
    "Mediocre product, expected better.",
    "Love everything about this!",
    "Complete waste of money.",
    "Good value for the price.",
    "Very satisfied, will buy again!",
    "Horrible experience from start to finish.",
    "It works as described.",
    "Outstanding quality and design!",
    "Disappointed with the results.",
    "Perfect! Exactly what I wanted.",
]

sentiments = [
    'Positive', 'Negative', 'Neutral', 'Positive', 'Negative',
    'Positive', 'Positive', 'Negative', 'Neutral', 'Positive',
    'Negative', 'Positive', 'Negative', 'Positive', 'Positive',
    'Negative', 'Neutral', 'Positive', 'Negative', 'Positive'
]

df = pd.DataFrame({'review': reviews_data, 'sentiment': sentiments})

print("Sample Reviews:")
print(df.head(10))

# 1. Lexicon-based Sentiment Analysis
from nltk.sentiment import SentimentIntensityAnalyzer
try:
    import nltk
    nltk.download('vader_lexicon', quiet=True)
    sia = SentimentIntensityAnalyzer()

    df['vader_scores'] = df['review'].apply(lambda x: sia.polarity_scores(x))
    df['vader_compound'] = df['vader_scores'].apply(lambda x: x['compound'])
    df['vader_sentiment'] = df['vader_compound'].apply(
        lambda x: 'Positive' if x > 0.05 else ('Negative' if x < -0.05 else 'Neutral')
    )

    print("\n1. VADER Sentiment Scores:")
    print(df[['review', 'vader_compound', 'vader_sentiment']].head())
except:
    print("NLTK not available, skipping VADER analysis")

# 2. Textblob Sentiment (alternative)
try:
    from textblob import TextBlob

    df['textblob_polarity'] = df['review'].apply(lambda x: TextBlob(x).sentiment.polarity)
    df['textblob_sentiment'] = df['textblob_polarity'].apply(
        lambda x: 'Positive' if x > 0.1 else ('Negative' if x < -0.1 else 'Neutral')
    )

    print("\n2. TextBlob Sentiment Scores:")
    print(df[['review', 'textblob_polarity', 'textblob_sentiment']].head())
except:
    print("TextBlob not available")

# 3. Feature Extraction for ML
vectorizer = TfidfVectorizer(max_features=100, stop_words='english')
X = vectorizer.fit_transform(df['review'])
y = df['sentiment']

print(f"\n3. Feature Matrix Shape: {X.shape}")
print(f"Features extracted: {len(vectorizer.get_feature_names_out())}")

# 4. Machine Learning Model
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3, random_state=42)

# Naive Bayes classifier
nb_model = MultinomialNB()
nb_model.fit(X_train, y_train)
y_pred = nb_model.predict(X_test)

print("\n4. Machine Learning Results:")
print(f"Accuracy: {accuracy_score(y_test, y_pred):.2f}")
print("\nClassification Report:")
print(classification_report(y_test, y_pred))

# 5. Sentiment Distribution
fig, axes = plt.subplots(2, 2, figsize=(14, 8))

# Distribution of sentiments
sentiment_counts = df['sentiment'