Sentiment Analysis Using FFBP Neural Network for Profit of Commercial Products in Industry

Abstract

From its inception in the early 2000s to the present, sentiment analysis and opinion mining have taken several twists and turns. People are increasingly using social media sites to voice their opinions and communicate with others who share common ideas, thanks to advancements in technology, smartphone and internet networks, and ease of access to these services. As a result, a vast volume of data has been generated on the internet, necessitating the need to analyse it Sentiment research aids various organisations in determining how customers perceive their goods and services, as well as what improvements are needed to enhance them. The paper uses an inbuilt python library called Text Blob to perform sentiment analysis, which is the classification of tweets into positive, negative, and neutral on views on a certain product, for three platforms: twitter, Facebook, and news websites. It also discusses how Artificial Neural Networks (ANN) provide a platform to perform sentiment analysis in a much easier and therefore less time consuming process.In this article, feed-forward back propagation neural networks (FFBPNN) were used to split the task into training data, and a min-max method has been used to measure the information and analysing the sentiment accuracy rate with ANN. To have a quantitative approach to the findings and quantify the success of ANN, precision, memory, and accuracy were measured. We discovered that this kind of neural network is very effective at accurately predicting the outcome.