The researchers explained that social media has become the dominant form of communication for people and companies looking to promote and sell their products and services. Correctly understanding and responding to customer reviews on Twitter, Facebook, and other social media platforms is critical to success, but it’s a time-consuming process.
Scientists plan to do this through sentiment analysis. This term refers to the automated process of identifying emotions – positive, negative, or neutral – associated with a text. Artificial intelligence refers to the logical analysis of data and responding to it. A team of scientists has developed a methodology that accurately identifies sarcasm in social media texts based on it.
The researchers taught the model to find patterns that often indicate sarcasm and combined them with the training program. To do this, they loaded a huge amount of data into the model, and then checked its accuracy.
“Sarcasm is not always easy to recognize in conversation, so you can imagine that it is even more difficult for a computer program to do it in a text. We have developed an interpretable deep learning model – it helps to identify important sarcastic words from the input text, and recurrent blocks study the relationships between these words in order to better classify the text, ”the scientists explained.
They added that sarcasm used to be a major obstacle to improving the accuracy of sentiment analysis, especially on social media, since sarcasm relies heavily on intonation, facial expressions and gestures that cannot be represented in the text. However, it turned out that the model can cope with this task as well.