Affective computing, the area of technology which the topic of this research paper falls under, is concerned with sensing the emotions of users and using the data gathered from the inputs that these users provide to serve them better. Speech Emotion Recognition, or SER for short has a growing number of applications, including intelligent toys and psychiatric diagnoses (Peerzade, Deshmukh, & Waghmare, 2018).
Despite the strides that have been made in this area of computing from its inception in 1996 (Schuller, 2018), evidence from the literature to be examined in this paper suggests that there is much room for improvement and if these improvements are made, SER could prove useful to recreational users and professionals across industries.
This research project will comparatively analyse various machine learning approaches to emotion recognition using speech, and to attempt to suggest optimal techniques for use throughout the development process to support the realisation of this goal.