Time based sentiment analysis using artificial neural networks and bert language model: Exploring comments on whatsapp's new privacy policy

Today, with the effect of developing social network culture and computer technologies, the satisfaction levels of millions of users about the product or service they can be measured. These datas are actively used in the fields of marketing strategies, advertising and customer support, and the risks and benefits ot the actions that organisations can take can be measured by analyzing these datas. Sentiment analysis is seen among decision support systems because it supports the decision to be taken. In this study, the time-based sentiment score change was analyzed by measuring the sentiment scores of the comments about the WhatsApp new privacy policy on Twitter with the Python programming language. The perception of Turkish text data is provided by using the BERTurk model, which is the developed version of the BERT language model, which is one of the pre-trained natural language processing models for the perception of the Turkish language. In order to classify the texts, a model with a multilayer perceptron, one of the artificial neural network models, was designed. Sentiment analysis was performed by obtaining tweets from Twitter regarding the WhatsApp new privacy policy, and the change in sentiment score over time was observed. Keywords: BERT, Multilayer Perceptron, Sentiment Analysis, Artificial Neural Networks.