MLOps Automation – Developing a RESTful API for Text Based Emotion Detection

MLOps Automation – Developing a RESTful API for Text Based Emotion Detection

연구 분야: Safety

학회: IC3-2022: Proceedings of the 2022 Fourteenth International Conference on Contemporary Computing

Identifying the emotional state of a person or group of people by analyzing their written works often appears to be challenging but also important. In many situations, we see that emotions from textual expressions cannot be detected directly using emotional words alone but also results from the evaluation of the conceptual meaning and interactions which are expressed in a written article. A person's emotion may be revealed by their facial expression, speech, or text. In todays’ world we have seen that efforts have been made in identifying emotion from speech and facial expressions, but the aspect of text-based emotion detection is still lagging. The detection of human emotions from text is becoming progressively important from an applicative perspective. The paper proposed an algorithm for detecting emotions from text using logistic regression. The authors have developed a RESTful API to serve as a mediator between a client device and the trained model by adopting the principles of Machine Learning Operations (MLOps).