Sentiment Analysis in VoIP messages using machne learning

VoIP devices are part of Internet-of-Things(IoT) devices that are accessible through the internet. VoIP has become a core technology to many businesses and general population across the world simplifying many communication processes by making it consumer friendly and globally accessible to all. It has been predicted that there will be 200billion call per minute in 2020 generating 86.20 billion in global revenues[1].

This project proposes to develop a software that converts VoIP messages to text using open source API and to perform sentiment analysis on such messages. This project addresses the issues in the QoS of the calls for example, poor quality calls, latency, reliability etc. The uttered keywords can be mapped to emotional behaviour of the user using a sentiment score using ML.

Training data can be used to look into the sentiment of the message and a visual sentiment ontology can be created using the emotions and the sentiment score. Using machine learning process the user’s emotional behaviour can be used to gauge opinions of individuals or groups and can be used to quantify emotions, attititudes related to the topic concerned. This emotional mapping can be used in various areas like social engineering, identifying emotionally vulnerable users.

1. https://voip.review/2019/02/20/predictions-voip-services-technologies-2019/