Publications

Abstract : In this work, an effective prediction of movie and music genre identification based on the emotions that are perceived by the viewers while watching media clips is achieved. The aim is to employ a heterogeneous ensemble technique that can relate the data to its respective genres and understand, which approach gives better results. By extracting the right features from the dataset upon using feature selection methods, better genre prediction is achieved. The dataset undergoes feature scaling and then plugged into different classifiers wherein it classifies the genres accordingly. The main contribution of the work is to identify media genres from brain signals such as EEG and MEG. In addition to it, different fusions of the brain data with multimedia data are validated and the corresponding best accuracy is computed.