@MASTERSTHESIS\{IMM2012-06384,
    author       = "M. Staszewski",
    title        = "Design and Implementation of Affective Android Music Player",
    year         = "2012",
    school       = "Technical University of Denmark, {DTU} Informatics, {E-}mail: reception@imm.dtu.dk",
    address      = "Asmussens Alle, Building 305, {DK-}2800 Kgs. Lyngby, Denmark",
    type         = "",
    note         = "Supervised by Associate Professor Ole Winther, owi@imm.dtu.dk, {DTU} Informatics",
    url          = "http://www.imm.dtu.dk/English.aspx",
    abstract     = "An affective load estimation model based on extended Affective Norms for English Words list [2] has been created. The model is used to assess emotional load (in terms of valence and arousal) of a corpus of lyrics. The {ANEW} [2] list has been extended with an algorithm using Latent Semantic Analysis [3] to compute term to term similarities which were then used to derive affective estimations for words not in the {ANEW} [2] list. Correlation between lyrics affective load and audio features is confirmed using Pearson measure to detect statistical significance. An Android music player is subsequently developed and audio parameters and lyrics affective load are used by the algorithm implemented in this application to assign a point in valence/arousal space for selected songs. The application allows the user to change assigned valence/arousal parameters. Bayesian classifier based model is trained to predict user preferences. An affective load estimation model based on extended Affective Norms for English Words list [1] has been created. The model is used to assess emotional load (in terms of valence and arousal) of a corpus of lyrics. The {ANEW} [1] list has been extended with an algorithm using Latent Semantic Analysis [2] to compute term to term similarities which were then used to derive affective estimations for words not in the {ANEW} [1] list. Correlation between lyrics affective load and audio features is confirmed using Pearson measure to detect statistical significance. An Android music player is subsequently developed and audio parameters and lyrics affective load are used by the algorithm implemented in this application to assign a point in valence/arousal space for selected songs. The application allows the user to change assigned valence/arousal parameters. Bayesian classifier based model is trained to predict user preferences."
}