Dark Mode
Turkish Music Emotion Analysis Dataset
Data Science and Analytics
Tags and Keywords
Trusted By
"No reviews yet"
Free
About
Categorises Turkish music into four distinct moods: happy, sad, angry, and relaxed. This dataset provides a robust foundation for research in Music Emotion Recognition (MER), featuring 400 audio samples from which 50 features were originally extracted. Through feature selection, a refined set of key audio characteristics has been identified, making it highly efficient for machine learning and audio analysis classification tasks. It is designed to unlock insights into the emotional content of music.
Columns
- Class: The target variable indicating the music's mood, with four unique values: happy, sad, angry, and relaxed.
- _Fluctuation_Mean: A measure of the average fluctuation in the audio signal.
- _MFCC_Mean_2: The mean value of the second Mel-frequency cepstral coefficient.
- _MFCC_Mean_4: The mean value of the fourth Mel-frequency cepstral coefficient.
- _MFCC_Mean_7: The mean value of the seventh Mel-frequency cepstral coefficient.
- _Roughness_Mean: The average sensory roughness of the audio.
- _Roughness_Slope: The rate of change of the audio's sensory roughness.
- _Zero-crossingrate_Mean: The mean rate at which the signal changes sign.
- _AttackTime_Mean: The average time for a sound to reach its peak amplitude.
- _Eventdensity_Mean: A measure of the average number of sound events per second.
- _Pulseclarity_Mean: The mean clarity of the underlying pulse in the music.
- _Chromagram_Mean_7: The mean value for the seventh chroma feature, relating to musical pitch.
- _HarmonicChangeDetectionFunction_Mean: The mean value from a function detecting changes in harmonic content.
- _HarmonicChangeDetectionFunction_PeriodAmp: The period amplitude from the harmonic change detection function.
Distribution
The dataset is provided in a single CSV file named
Turkish_Music_Mood_Recognition.csv, with a size of 35.41 kB. It is structured with 400 records and 14 columns.Usage
This dataset is highly suitable for developing and testing various audio analysis and machine learning models. Ideal applications include:
- Creating music recommendation systems based on user mood.
- Generating emotion-driven playlists automatically.
- Conducting research on audio feature analysis and classification.
- Building AI systems capable of recognising emotions from sound.
Coverage
The dataset focuses on Turkish music. It contains an equal distribution of samples across the four emotional classes, providing a balanced resource for classification tasks.
License
Attribution 4.0 International (CC BY 4.0)
Who Can Use It
- Machine Learning Enthusiasts: For building and training models for multi-class classification tasks.
- Audio Researchers: To analyse the relationship between audio features and perceived emotion.
- Musicologists: For studying the acoustic properties that define emotional expression in Turkish music.
- Data Scientists: To explore feature selection techniques and develop predictive audio models.
Dataset Name Suggestions
- Turkish Music Mood and Audio Features
- Emotion Recognition in Turkish Music
- Acoustic Features for Music Mood Classification
- Turkish Music Emotion Analysis Dataset
Attributes
Original Data Source:Turkish Music Emotion Analysis Dataset
QUALITY