Music Mood Classification using Python
€10-15 EUR
Betales ved levering
Automatic recognition of emotions in musical audio has gained increasing
attention in the field of music information retrieval (MIR) during the past few
years. The development in the field has coincided with the need for managing
large collections of digital audio for the public via web services such as Spotify
and Last.fm. This is reflected, for example, in the increasing number of
submitted systems in the annual Audio Music Mood Classification (AMC)
contest part of the Music Information Retrieval Evaluation eXchange3
(MIREX). The substantial variance in the submitted systems in the contest
and in the approaches in the MIR field in general indicates a lack of consensus
concerning the choice of an underlying psychological model for emotions or
moods in music and the establishment of precise machine learning conventions.
Despite research in musicology studying the influence of specific cues in the
musical structure on emotional expressions, there is no complete analytical
consensus about the required “ingredients”—i.e., acoustical features extracted
from music—to build the optimal models for emotion recognition, in the limits
imposed by the subjectivity of [login to view URL] descriptors for each music track
provided for this activity are the state-of-the-art descriptors used in this area
(see table on next page for details). Your task is to determine which sets of
features are most useful for learning the accurate classifiers. In particular your
task isto explore different feature sets and machine learning algorithms with the
aim to obtain the most accurate classifier. As part of your exploration you are
supposed to get the corresponding accuracies for several algorithms (e.g.
SVM, KNN, Decision Trees, Neural Networks, Ensemble, Naive Bayes), and
applying training set, bootstrap, and cross-validation methods to get the
training and testing accuracies. What kind of features (timbre, rhythm, pitch,
dynamics, structure or harmony) are more relevant for predicting emotion in
music in the provided data set?
Prosjekt-ID: #22309917
Om prosjektet
Tildelt til:
Note: I read all of your requirements and i can deliver you with quality work. I have been worked and implemented all the techniques which you have mentioned to be done. Greetings! Welcome to DEEP LEARNING, MACHINE LE Mer