Spam-Classification

• Importing necessary libraries like numpy, pandas, matplotlib, seaborn, and warnings.
• Reading the 'spam.csv' file containing SMS messages and labels as spam or ham using pandas read_csv function.
• Dropping the null values from the dataset.
• Renaming the column names as 'label' and 'message.'
• Creating a count plot to visualize the imbalance in the dataset between spam and ham messages.
• Handling the imbalanced dataset by oversampling spam messages in the dataset using for loop.
• Creating another count plot to visualize the balance between spam and ham messages after oversampling.
• Creating a new feature called 'word_count' containing the number of words in each SMS message.
• Plotting a distribution plot for word count for spam and ham messages separately.
• Creating another feature called 'contain_currency_symbol' to check if an SMS message contains a currency symbol or not.
• Creating a count plot to visualize the number of SMS messages containing currency symbols for spam and ham messages separately.
• Creating another feature called 'contain_number' to check if an SMS message contains a number or not.
• Creating a count plot to visualize the number of SMS messages containing numbers for spam and ham messages separately.
• Importing necessary libraries like nltk, re, and stopwords.
• Cleaning the SMS messages using the WordNetLemmatizer, removing special characters, converting them to lowercase, tokenizing them, removing stop words, and lemmatizing them.
• Creating a bag of words model using the TfidfVectorizer, extracting independent and dependent variables from the dataset.
• Splitting the dataset into training and testing sets using the train_test_split function.
• Fitting the Naive Bayes model to the training set and calculating the average F1 score and standard deviation using the cross_val_score function.
• Generating a classification report and confusion matrix for the Naive Bayes model using the classification_report and confusion_matrix functions, respectively.