Comparative Analysis of Machine Learning Techniques for Detecting Sentiments in Social Media

This paper provides an extensive discussion of the machine learning algorithms applied to the sentiment analysis on the social media, involving the use of the Naive Bayes, Support Vector Machines (SVM), and Deep learning models evaluation and comparison. The growth of the social network content at the rates exponential, computerized assessment and interpretation of the valuer of the client have become extremely important to serve the areas of the market researches from the social opinion monitoring. We conducted a systematic study on the viability of various machine learning methodologies that aim to deal with the complicated characteristics as well as peculiar nature of Language Processing Natural (LPN), which are most-likely caused by the enormous amount of data on social media platforms. According to our results, the Deep Learning based models incorporate more complex neural network structure and thus, CNNs and RNNs are more likely to outperform other explanations. The critical point of their success lies in the power of models to capture the semantic and contextual nature of language. Nevertheless, the exploration also outlines the computational needs of Deep Learning methods, which imply some critical requirements for their use in the contemporary sentiment analysis applications. We cover issues of processing speed and discussion of trade-offs in terms of classification accuracy versus computational efficiency, thus point out the implementation problems that scaled adoption of the models brings.