Scribd
Upload
2 views

TopicClassifierbyDavidCaleb

The document outlines a Python script for text classification using the 20 Newsgroups dataset, employing various machine learning models including Naive Bayes, SVM, Random Forest, and Logistic Regression. It includes data preprocessing steps, feature extraction with TF-IDF, model training, hyperparameter tuning, and evaluation through accuracy scores and confusion matrices. The best model is saved for future use, demonstrating a comprehensive approach to text classification tasks.

Uploaded by

wet42701
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
2 views

TopicClassifierbyDavidCaleb

The document outlines a Python script for text classification using the 20 Newsgroups dataset, employing various machine learning models including Naive Bayes, SVM, Random Forest, and Logistic Regression. It includes data preprocessing steps, feature extraction with TF-IDF, model training, hyperparameter tuning, and evaluation through accuracy scores and confusion matrices. The best model is saved for future use, demonstrating a comprehensive approach to text classification tasks.

Uploaded by

wet42701
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 7

# Import necessary libraries

import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns
from sklearn.datasets import fetch_20newsgroups
from sklearn.feature_extraction.text import TfidfVectorizer
from sklearn.model_selection import train_test_split, GridSearchCV,
cross_val_score
from sklearn.naive_bayes import MultinomialNB
from sklearn.svm import SVC
from sklearn.ensemble import RandomForestClassifier
from sklearn.linear_model import LogisticRegression
from sklearn.metrics import classification_report, confusion_matrix,
accuracy_score, roc_curve, auc
from sklearn.pipeline import Pipeline
from nltk.stem import WordNetLemmatizer
from nltk.tokenize import word_tokenize
import nltk
import string

# Download NLTK data (only required once)


nltk.download('punkt')
nltk.download('wordnet')

[nltk_data] Downloading package punkt to


[nltk_data] C:\Users\Usuario\AppData\Roaming\nltk_data...
[nltk_data] Package punkt is already up-to-date!
[nltk_data] Downloading package wordnet to
[nltk_data] C:\Users\Usuario\AppData\Roaming\nltk_data...

True

# Load the dataset


categories = ['sci.space', 'comp.graphics', 'talk.politics.guns',
'rec.sport.baseball', 'sci.med']
newsgroups = fetch_20newsgroups(subset='all', categories=categories,
shuffle=True, random_state=42)

# Figure adjustments
fig = plt.figure(figsize=(12, 15)) # Larger size for better
visualization
gs = fig.add_gridspec(2, 2)
gs.update(wspace=0.4, hspace=0.7) # Increases space between subplots

# Background and color settings


background_color = "#f8fafc"
color_palette = ["#94a3b8", "#64748b", "#475569", "#334155",
"#1e293b"]
fig.patch.set_facecolor(background_color)
axes = [fig.add_subplot(gs[i, j]) for i in range(2) for j in range(2)]
for ax in axes:
ax.set_facecolor(background_color)

# Title of the plot


axes[0].spines["bottom"].set_visible(False)
axes[0].spines["left"].set_visible(False)
axes[0].spines["top"].set_visible(False)
axes[0].spines["right"].set_visible(False)
axes[0].tick_params(left=False, bottom=False)
axes[0].set_xticklabels([])
axes[0].set_yticklabels([])
axes[0].text(0.5, 0.5,
'Topic Distribution and\nDocument Analysis\
n_________________',
horizontalalignment='center',
verticalalignment='center',
fontsize=18, fontweight='bold',
fontfamily='serif',
color="#000000")

# Plot 1: Distribution of Topics


ax1 = axes[1]
ax1.set_title("Distribution of Topics", loc='center', fontsize=14,
fontweight='bold', fontfamily="serif", color="#000000")
sns.countplot(x=newsgroups.target, palette=color_palette, ax=ax1)
ax1.set_xticks(np.arange(len(categories)))
ax1.set_xticklabels(newsgroups.target_names, rotation=30, ha='right',
fontsize=10)
ax1.set_xlabel("")
ax1.set_ylabel("Count", fontsize=12)

# Plot 2: Document Length Distribution


ax2 = axes[2]
ax2.set_title("Document Length Distribution", loc='center',
fontsize=14, fontweight='bold', fontfamily="serif", color="#000000")
doc_lengths = [len(doc.split()) for doc in newsgroups.data]
sns.histplot(doc_lengths, bins=50, kde=True, color=color_palette[2],
ax=ax2)
ax2.set_xlabel("Number of Words", fontsize=12)
ax2.set_ylabel("Frequency", fontsize=12)

# Plot 3: Most Common Words (Optional)


from sklearn.feature_extraction.text import CountVectorizer

ax3 = axes[3]
ax3.set_title("Most Common Words", loc='center', fontsize=14,
fontweight='bold', fontfamily="serif", color="#000000")
vectorizer = CountVectorizer(stop_words='english', max_features=20)
word_counts = vectorizer.fit_transform(newsgroups.data)
word_freq = dict(zip(vectorizer.get_feature_names_out(),
word_counts.sum(axis=0).tolist()[0]))
sns.barplot(x=list(word_freq.values()), y=list(word_freq.keys()),
palette=color_palette, ax=ax3)
ax3.set_xlabel("Frequency", fontsize=12)
ax3.set_ylabel("Words", fontsize=12)

# Adjust and show


plt.show()

C:\Users\Usuario\AppData\Local\Temp\ipykernel_9492\3698852670.py:33:
FutureWarning:

Passing `palette` without assigning `hue` is deprecated and will be


removed in v0.14.0. Assign the `x` variable to `hue` and set
`legend=False` for the same effect.

sns.countplot(x=newsgroups.target, palette=color_palette, ax=ax1)


C:\Users\Usuario\AppData\Local\Temp\ipykernel_9492\3698852670.py:55:
FutureWarning:

Passing `palette` without assigning `hue` is deprecated and will be


removed in v0.14.0. Assign the `y` variable to `hue` and set
`legend=False` for the same effect.

sns.barplot(x=list(word_freq.values()), y=list(word_freq.keys()),
palette=color_palette, ax=ax3)
C:\Users\Usuario\AppData\Local\Temp\ipykernel_9492\3698852670.py:55:
UserWarning:
The palette list has fewer values (5) than needed (20) and will cycle,
which may produce an uninterpretable plot.
sns.barplot(x=list(word_freq.values()), y=list(word_freq.keys()),
palette=color_palette, ax=ax3)
# Advanced Text Preprocessing
# Define a function to clean and lemmatize text
def preprocess_text(text):
# Remove punctuation
text = text.translate(str.maketrans('', '', string.punctuation))
# Tokenize text
tokens = word_tokenize(text.lower())
# Lemmatize tokens
lemmatizer = WordNetLemmatizer()
tokens = [lemmatizer.lemmatize(token) for token in tokens]
return ' '.join(tokens)

# Apply preprocessing to the dataset


preprocessed_data = [preprocess_text(doc) for doc in newsgroups.data]

# Feature Extraction with TF-IDF


vectorizer = TfidfVectorizer(stop_words='english', max_df=0.7,
ngram_range=(1, 2))
X = vectorizer.fit_transform(preprocessed_data)
y = newsgroups.target

# Split the dataset into training and testing sets


X_train, X_test, y_train, y_test = train_test_split(X, y,
test_size=0.3, random_state=42)

# Compare Multiple Models


models = {
'Naive Bayes': MultinomialNB(),
'SVM': SVC(kernel='linear', probability=True, random_state=42),
'Random Forest': RandomForestClassifier(n_estimators=100,
random_state=42),
'Logistic Regression': LogisticRegression(max_iter=1000,
random_state=42)
}

results = {}
for name, model in models.items():
# Train the model
model.fit(X_train, y_train)
# Make predictions
y_pred = model.predict(X_test)
# Evaluate the model
accuracy = accuracy_score(y_test, y_pred)
results[name] = accuracy
print(f"{name} Accuracy: {accuracy:.4f}")

Naive Bayes Accuracy: 0.9808


SVM Accuracy: 0.9787
Random Forest Accuracy: 0.9238
Logistic Regression Accuracy: 0.9780

# Hyperparameter Tuning with Grid Search (Example: SVM)


param_grid = {'C': [0.1, 1, 10], 'kernel': ['linear', 'rbf']}
grid_search = GridSearchCV(SVC(probability=True, random_state=42),
param_grid, cv=3)
grid_search.fit(X_train, y_train)
print(f"Best SVM Parameters: {grid_search.best_params_}")

Best SVM Parameters: {'C': 10, 'kernel': 'linear'}

# Cross-Validation
cv_scores = cross_val_score(grid_search.best_estimator_, X, y, cv=5)
print(f"Cross-Validation Accuracy: {np.mean(cv_scores):.4f}")

Cross-Validation Accuracy: 0.9794

# Visualize Results
# Confusion Matrix for the best model
best_model = grid_search.best_estimator_
y_pred = best_model.predict(X_test)
cm = confusion_matrix(y_test, y_pred)
plt.figure(figsize=(8, 6))
sns.heatmap(cm, annot=True, fmt='d', cmap='Blues',
xticklabels=newsgroups.target_names,
yticklabels=newsgroups.target_names)
plt.title('Confusion Matrix')
plt.xlabel('Predicted')
plt.ylabel('Actual')
plt.show()
# Save the Model
import joblib
joblib.dump(best_model, 'text_classification_model.pkl')

['text_classification_model.pkl']

You might also like