Enhancement of Twitter Spam Detection Using Naïve Bayes Algorithm

Abstract

The Naive Bayes classification algorithm is a widely used method suitable for both binary and multiclass classification tasks. Unlike numerical input variables, Naive Bayes performs well when dealing with categorical input variables. It is commonly employed in applications such as sentiment analysis, spam filtering, and recommendation systems. One advantage of Naive Bayes is its ability to make predictions and anticipate data based on past outcomes. It is known for its simplicity and efficiency, requiring less training data compared to other models. However, a major limitation is its assumption of independent predictors, which may not hold true in real-world scenarios. Despite this drawback, Naive Bayes exhibits better performance and offers a wider range of predictions when researchers incorporate improvements and advancements. This makes it a suitable choice for multi-class prediction problems. Researchers have conducted extensive testing and simulations, leading to significant advancements in the algorithm's performance, including improvements in vocabulary, accuracy, and speed. Nonetheless, there are still unresolved challenges in automatic text processing, particularly in the domain of spam identification in social networks. Ongoing research aims to overcome these challenges and further enhance the capabilities of Naive Bayes and automatic text processing techniques.