Spam Email Classification

A Machine Learning project for detecting Spam and Ham messages using TF-IDF and Logistic Regression.

---

! ✓

Overview

This project uses Natural Language Processing (NLP) and Machine Learning to classify messages into two categories:

• Ham — Normal / legitimate messages
• Spam — Unwanted or suspicious messages

Machine Learning Pipeline

Text Data
   ↓
Data Cleaning
   ↓
Label Encoding
   ↓
TF-IDF Feature Extraction
   ↓
Train/Test Split
   ↓
Logistic Regression
   ↓
Prediction
   ↓
Model Evaluation

---

Dataset

The dataset contains 6,464 messages with two columns:

Column	Description
label	Target category: spam or ham
text	Message content

Example

label	text
ham	Hello, how are you?
spam	Congratulations! You won a free prize!

---

Technologies Used

• Python
• Pandas
• NumPy
• Scikit-learn
• TF-IDF
• Matplotlib
• Seaborn
• Joblib
• Flask (optional for deployment)

---

Machine Learning Workflow

1️ Data Loading

import pandas as pd

email = pd.read_csv("after data cleaning email.csv")

Data Cleaning

email["text"] = email["text"].fillna("empty message")

Label Encoding

from sklearn.preprocessing import LabelEncoder

encoder = LabelEncoder()
encoded = encoder.fit_transform(email["label"])

TF-IDF Feature Engineering

from sklearn.feature_extraction.text import TfidfVectorizer

vectorizer = TfidfVectorizer()

X_tfidf = vectorizer.fit_transform(email["text"])

Train/Test Split

from sklearn.model_selection import train_test_split

X_train, X_test, y_train, y_test = train_test_split(
    X_tfidf,
    encoded,
    test_size=0.2,
    random_state=42,
    stratify=encoded
)

Model Training

from sklearn.linear_model import LogisticRegression

model = LogisticRegression(max_iter=1000)

model.fit(X_train, y_train)

Prediction

predictions = model.predict(X_test)

Prediction probability:

probabilities = model.predict_proba(X_test)

Model Evaluation

from sklearn.metrics import accuracy_score

accuracy = accuracy_score(y_test, predictions)

print("Accuracy:", accuracy)

---

Saving the Model

After training, save both the trained model and TF-IDF vectorizer:

import joblib

joblib.dump(model, "spam_model.pkl")
joblib.dump(vectorizer, "tfidf_vectorizer.pkl")

Saved Files

spam_model.pkl
tfidf_vectorizer.pkl

This allows the trained model to be reused without training it again.

---

Example Prediction

message = ["Congratulations! You have won a free prize"]

message_tfidf = vectorizer.transform(message)

prediction = model.predict(message_tfidf)

print(prediction)

---

Project Objectives

• Clean and preprocess text data
• Encode categorical labels
• Apply TF-IDF feature engineering
• Split data into training and testing sets
• Train a Logistic Regression classifier
• Predict Spam and Ham messages
• Calculate model accuracy
• Save the trained model
• Save the TF-IDF vectorizer
• Prepare the model for deployment

---

Concepts Demonstrated

• Data Cleaning
• Label Encoding
• Natural Language Processing
• TF-IDF
• Feature Engineering
• Train/Test Split
• Supervised Learning
• Logistic Regression
• Classification
• Model Training
• Prediction
• Prediction Probability
• Accuracy Score
• Model Saving with Joblib

---

Future Improvements

• Build a Flask web application
• Add a message input form
• Display prediction probability
• Improve text preprocessing
• Compare different classification algorithms
• Add confusion matrix and classification report
• Deploy the application online

---

Installation

Clone the repository:

git clone https://github.com/rabia743/Email-Spam-detector-ML-model.git
cd Email-Spam-detector-ML-model

Install required libraries:

pip install pandas numpy scikit-learn matplotlib seaborn scipy joblib

---

Important

The same TF-IDF vectorizer used during training must be reused during prediction.

Training:

Text
 ↓
TF-IDF Vectorizer
 ↓
Machine Learning Model

Prediction:

New Text
 ↓
Same TF-IDF Vectorizer
 ↓
Saved Model
 ↓
Prediction

---

Author

Rabia Zafar

Python Developer | Machine Learning Developer

---

If you found this project useful, consider giving it a star!

Built with Python, NLP & Machine Learning

---

💌 Thanks for Visiting! 💌