Variable Selection and Interpretation of Covariance Principal Components

Al-Kandari, Noriah M.; Jolliffe, Ian T.

doi:10.1081/sac-100002371

Cited by 48 publications

(27 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This is because the original features are more straightforward to explain than the PCs, as PCs unfortunately require a linear combination of the whole initial matrix to account for themselves. So the selection of optimal individual features (representing the nine PCs) was performed instead using feature selection principles as outlined by Jolliffe (see Al-Kandari & Jolliffe, 2001). In other words, the highest correlating feature for each PC was chosen and highlighted in Table 2.…”

Section: Feature Reductionmentioning

confidence: 99%

Are the Emotions Expressed in Music Genre-specific? An Audio-based Evaluation of Datasets Spanning Classical, Film, Pop and Mixed Genres

Eerola

2011

Journal of New Music Research

View full text Add to dashboard Cite

Empirical studies of emotions in music have described the role of individual musical features in recognizing particular emotions. However, no attempts have been made as yet to establish if there is a link between particular emotions and a specific genre. Here this is investigated by analysing nine separate datasets that represent categories ranging from classical (three sets), and film music (two), to popular music (two), and mixed genre (two). A total of 39 musical features were extracted from the audio. Models were then constructed from these to explain self-reports of valence and arousal, by using multiple and Random Forest regression. The models were fully validated across the datasets, suggesting low generalizability between the genres for valence (16% variance was accounted for) and moderately good generalizability between the genres for arousal (43%). In contrast, the generalizability within genres was considerably higher (43% and 62% respectively), which suggests that emotions, especially those that express valence, operate differently depending on the musical genre. The most reliable musical features of affects across genres were identified, yielding a ranked set of features most likely to operate across the genre. In conclusion, the implications of the findings, and the genre-specificity of emotions in music are discussed.

show abstract

Section: Feature Reductionmentioning

confidence: 99%

Are the Emotions Expressed in Music Genre-specific? An Audio-based Evaluation of Datasets Spanning Classical, Film, Pop and Mixed Genres

Eerola

2011

Journal of New Music Research

View full text Add to dashboard Cite

show abstract

“…Their aim is to eliminate possible incommensurability of subjects within the individual data sets (data scaling) and size diferences between data sets (coniguration scaling), see (Gower and Dijksterhuis 2004). Basically, translation, rotation, and dilation, which performed in the respected order, are the kinds of transformations that may be deemed desirable before embarking on the actual procrustes matching (Digby and Kempton 1987;Al Kandari and Jollife 2001;Bakhtiar and Siswadi 2011). PA can also be utilized to determine the goodness of it between a data matrix and its approximation .…”

Section: Introductionmentioning

confidence: 99%

RETRACTED ARTICLE: Subject choice in educational data sets by using principal component and procrustes analysis

Khan

Jana

Bera

et al. 2016

Model. Earth Syst. Environ.

View full text Add to dashboard Cite

“…This is because they are more straightforward to explain than the PCs, which are less so, as a linear combination of the whole initial matrix is needed to account for them. Therefore, the selection of optimal individual features representing the seven PCs was carried out using feature selection principles as outlined by Jolliffe (see Al-Kandari & Jolliffe, 2001). In other words, the highest correlating feature for each PC was chosen (these correlations are marked in bold in Table 3).…”

mentioning

confidence: 99%

Timbre and Affect Dimensions: Evidence from Affect and Similarity Ratings and Acoustic Correlates of Isolated Instrument Sounds

2012

View full text Add to dashboard Cite

considerable effort has been made towards understanding how acoustic and structural features contribute to emotional expression in music, but relatively little attention has been paid to the role of timbre in this process. Our aim was to investigate the role of timbre in the perception of affect dimensions in isolated musical sounds, by way of three behavioral experiments. In Experiment 1, participants evaluated perceived affects of 110 instrument sounds that were equal in duration, pitch, and dynamics using a three-dimensional affect model (valence, energy arousal, and tension arousal) and preference and emotional intensity. In Experiment 2, an emotional dissimilarity task was applied to a subset of the instrument sounds used in Experiment 1 to better reveal the underlying affect structure. In Experiment 3, the perceived affect dimensions as well as preference and intensity of a new set of 105 instrument sounds were rated by participants. These sounds were also uniform in pitch, duration, and playback dynamics but contained systematic manipulations in the dynamics of sound production, articulation, and ratio of high-frequency to low-frequency energy. The affect dimensions for all the experiments were then explained in terms of the three kinds of acoustic features extracted: spectral (e.g., ratio of high-frequency to low-frequency energy), temporal (e.g., attack slope), and spectro-temporal (e.g., spectral flux). High agreement among the participants' ratings across the experiments suggested that even isolated instrument sounds contain cues that indicate affective expression, and these are recognized as such by the listeners. A dominant portion (50-57%) of the two dimensions of affect (valence and energy arousal) could be predicted by linear combinations of few acoustic features such as ratio of high-frequency to low-frequency energy, attack slope, and spectral regularity. Links between these features and those observed in the vocal expression of affects and other sound phenomena are discussed.

show abstract

Variable Selection and Interpretation of Covariance Principal Components

Cited by 48 publications

References 9 publications

Are the Emotions Expressed in Music Genre-specific? An Audio-based Evaluation of Datasets Spanning Classical, Film, Pop and Mixed Genres

Are the Emotions Expressed in Music Genre-specific? An Audio-based Evaluation of Datasets Spanning Classical, Film, Pop and Mixed Genres

RETRACTED ARTICLE: Subject choice in educational data sets by using principal component and procrustes analysis

Timbre and Affect Dimensions: Evidence from Affect and Similarity Ratings and Acoustic Correlates of Isolated Instrument Sounds

Contact Info

Product

Resources

About