Email/ SMS Spam Classification

Kaggle Dataset - https://www.kaggle.com/datasets/uciml/sms-spam-collection-dataset

The SMS Spam Collection is a set of SMS-tagged messages that have been collected for SMS Spam research. It contains one set of SMS messages in English of 5,574 messages, tagged according to being ham (legitimate) or spam.

The following tasks were performed in the sms-spam-detection.ipynb file

1. Data cleaning

• Dropped the column null values
• Dropped for duplicate rows
• Encoded the target variable

2. Exploratory Data Analysis

• Visualized Spam and Not Spam dataset and found the data was imbalanced
• NLTK Punkt sentence tokenizer is used to calculate the number of characters, number of words, and number of sentences in the text feature.
• After adding the above three features, the dataset is visualized. The correlation Matrix shows a high correlation between the three new features. Therefore, only one feature among them is used for further pre-processing tasks.
• Spam and Not Spam words are visualized using word cloud

3. Text Preprocessing

The following pre-processing tasks were implemented:-

• Lower case
• Tokenization
• Removing special characters
• Removing stop words and punctuation
• Stemming

4. Model building

• CountVectorizer and TfidfVectorizer are used to convert text to vector values
• Naive Bayes (GaussianNB,MultinomialNB,BernoulliNB)
• LogisticRegression
• Support Vector Classifier
• KNeighbors Classifier
• Decision Tree Classifier
• Random Forest Classifier
• AdaBoost Classifier
• Bagging Classifier
• Extra Trees Classifier
• Gradient Boosting Classifier
• XGBoost Classifier
• Voting Classifier
• Stacking

5. Website

• Designed a webpage using Streamlit to deploy the model