Real-Time Twitter Spam Detection and Sentiment Analysis using Machine Learning and Deep Learning Techniques.

Area of Science:

• Natural Language Processing
• Machine Learning
• Social Media Analysis

Background:

• Social media platforms face significant challenges from spam accounts, which compromise user experience through malicious links and repetitive content.
• Effective spam detection and sentiment analysis are crucial for maintaining platform integrity and understanding user engagement.

Purpose of the Study:

• To develop a robust system for classifying tweets as either 'spam' or 'ham' (legitimate).
• To evaluate the sentiment or emotion expressed within tweets.
• To compare the performance of various machine learning classifiers for both spam detection and sentiment analysis.

Main Methods:

• Tweet preprocessing and feature extraction were performed.
• Spam detection utilized classifiers such as Decision Tree, Logistic Regression, Naïve Bayes (Multinomial and Bernoulli), Support Vector Machine, and Random Forest.
• Sentiment analysis employed Stochastic Gradient Descent, Support Vector Machine, Logistic Regression, Random Forest, Naïve Bayes, and deep learning models including Recurrent Neural Network (RNN), Long Short-Term Memory (LSTM), Bidirectional LSTM (BiLSTM), and 1D Convolutional Neural Network (CNN).

Main Results:

• Extracted tweet features proved effective for accurately identifying spam content.
• The study analyzed and compared the performance of multiple classification algorithms for both tasks.
• A learning model was successfully created to associate tweets with specific sentiments.

Conclusions:

• The proposed system demonstrates satisfactory performance in distinguishing spam from legitimate tweets.
• Machine learning approaches, including deep learning, are effective for analyzing tweet sentiment.
• This research contributes to developing more secure and insightful social media platforms.