NdV_Code_By_RibkaA_Ass_7.py

[ribkaaramalla322]

This project builds a spam detection classification model using supervised learning techniques. The dataset consists of labeled SMS messages (spam or ham) and is preprocessed by encoding labels and vectorizing text using CountVectorizer. The dataset is split into training and testing sets (80/20) to evaluate generalization. Two models are trained: Logistic Regression and Naive Bayes, and their performances are compared. Evaluation metrics include accuracy, precision, recall, F1-score, confusion matrix, and ROC curve. The Logistic Regression model's feature importance is visualized to interpret the most influential words in spam classification. Confusion matrices are plotted to compare predicted vs actual outcomes. The ROC curve visually confirms the model’s predictive power. This project demonstrates effective text classification and model explainability using machine learning.