Author : Mr. Vikram Bhalerao and Mrs.Rushali Deshmukh

As the use of the internet is increasing, individuals are connected to each other using social media platforms like text messages, Facebook, Twitter, etc. This has led to the extent within the unfold of unsought messages far-famed as spam that is employed for selling, aggregation of personal data, or just to offend the individuals. Therefore, it’s crucial to own a powerful spam detection design that might stop these styles of messages. Spam detection in hissing platform like Twitter remains a tangle, thanks to short text and high variability within the language utilized in social media. In this paper, we tend to propose a CNN algorithmic technique and compare results with variants of CNN and with boosting algorithms. The model is supported by introducing the linguistics data in the illustration of the words with the assistance of knowledge-bases such as Word2vec and FastText. The use of these knowledge-bases improves the performance, by providing higher linguistics vector illustration of input testing words. Projected Experimental results with benchmark datasets, shows the effectiveness of the proposed approach with relevance to the accuracy, F1-score and response time.

Keywords: Convolutional Neural Network, Sentiment Analysis, Word2Vec, FastText.