COMPANY: CODTECH IT SOLUTIONS
NAME: SIDDHI TANAJI SATAV
INTERN ID: CTIS0158
DOMAIN: Python Programming
DURATION: 4 WEEEKS
MENTOR: NEELA SANTOSH
Internship Project Submission
CODTECH Internship – Task 4 Title: Spam Message Classification Using Scikit‑Learn
-
Objective Create a predictive model using scikit‑learn to classify text messages as spam or ham (not spam). The implementation and evaluation are presented in a Jupyter notebook, fulfilling the requirement to build and document a predictive model.
-
Tech Stack Language: Python 3 Environment: Jupyter Notebook (.ipynb) Libraries: pandas – data loading and preprocessing scikit-learn – machine learning (model, train–test split, metrics) numpy – numerical support (used by scikit‑learn) Install dependencies: bash pip install pandas scikit-learn numpy
-
Dataset File: spam.csv (placed in the same folder as the notebook). Structure: text label,message spam,"Win a free iPhone now! Click this link to claim your prize." ham,"Are we still meeting for lunch tomorrow at 1 pm?" ... label – target variable (spam or ham). message – SMS/email‑like text content. This small dataset is enough to demonstrate the full ML workflow; it can be replaced with a larger public spam dataset if needed.
-
Notebook Workflow Filename: e.g., Task4_SpamClassifier.ipynb
4.1. Data Loading and Exploration python import pandas as pd df = pd.read_csv("spam.csv") df.head() df['label'].value_counts() Loads data into a DataFrame. Checks class distribution to see how many spam vs ham messages exist.
4.2. Train–Test Split python from sklearn.model_selection import train_test_split X = df['message'] y = df['label']
X_train, X_test, y_train, y_test = train_test_split( X, y, test_size=0.2, random_state=42, stratify=y ) Splits the dataset into training and test sets to evaluate generalization.
4.3. Text Vectorization (TF‑IDF) python from sklearn.feature_extraction.text import TfidfVectorizer vectorizer = TfidfVectorizer(stop_words='english') X_train_vec = vectorizer.fit_transform(X_train) X_test_vec = vectorizer.transform(X_test) Converts raw text into numerical feature vectors using TF‑IDF, a common technique in spam detection.
4.4. Model Training python from sklearn.naive_bayes import MultinomialNB model = MultinomialNB() model.fit(X_train_vec, y_train) Uses Multinomial Naive Bayes, a standard algorithm for text classification, especially spam filters.
4.5. Evaluation python from sklearn.metrics import accuracy_score, confusion_matrix, classification_report y_pred = model.predict(X_test_vec) print("Accuracy:", accuracy_score(y_test, y_pred)) print("\nConfusion matrix:\n", confusion_matrix(y_test, y_pred)) print("\nClassification report:\n", classification_report(y_test, y_pred)) Reports accuracy, confusion matrix, precision, recall, and F1‑score for spam and ham. In markdown cells, the notebook explains what these metrics mean and comments on model performance. 4.6. Predicting New Messages python def predict_message(text): vec = vectorizer.transform([text]) return model.predict(vec)[0]
print(predict_message("Congratulations! You have won a free lottery. Click here.")) print(predict_message("Please send the project files by evening.")) Demonstrates how to use the trained model for new, unseen messages. Shows example outputs to illustrate practical use.
-
How to Run Ensure the following files are in the same folder: Task4_SpamClassifier.ipynb spam.csv Install dependencies (once): bash pip install pandas scikit-learn numpy Launch Jupyter: bash jupyter notebook Open Task4_SpamClassifier.ipynb and run all cells in order (Kernel → Restart & Run All). You should see: Data preview and label counts. Successful model training. Accuracy and other metrics printed. Example predictions for custom messages.
-
Deliverable Summary Jupyter Notebook: complete implementation and evaluation of the spam classifier. Dataset: spam.csv used for training and testing. The notebook clearly demonstrates the end‑to‑end ML pipeline: data loading, preprocessing, model building, evaluation, and prediction on new data using scikit‑learn.
Output
