Sentiment Analysis on Financial News Using Deep Learning Algorithm

Abstract

Sentiment analysis is the technique of computationally figuring out and categorizing reviews or comments expressed in a bit of textual content, especially a good way to decide whether or not the writer's mind-set in the direction and also very helpful to identify the customer’s opinion about the particular product or content. It is one of the active and wanted research areas in natural language processing. In existing work, machine learning algorithms like Support vector machine, Logistic regression, and Random forest have been used to classify the text from various sources like newspapers, reviews, social media comments, and predict the semantic orientations like Positive, Negative or Neutral. In that, accuracy is unsatisfied compared to the expectations. In this proposed work, both Long Short-Term Memory and Support Vector Machine are used for the classification and prediction of the sentiments from the financial news corpus. And the proposed Deep Learning algorithm is implemented in two ways, that is with pre-processing and without pre-processing. With pre-processing included stop words removal, punctuation removal, lower case conversion and tokenization. It performed better than without pre-processing. Finally, by comparing machine learning and deep learning algorithm, deep learning algorithm assures accuracy that is higher than machine learning algorithm.