A diagram on preliminary ideas for data generation during initial stages of my PhD on Music Composition at the QLD Con. The proccess start with a series of precomposed themes that will be feed into tyhe system as audio files. Individual will listen to each them and during this process several sources of data will be explored – i.e.
(i) Mental State Examination and Diferential Diagnosis;
(ii) Selected physiological parameters traditional known to correlate with mental state changes, i.e. skin impedance, pulse rate, respiratory rate, blood pressure and body temperature. During this process of data acquisition some individual demographics will be recorded too, i.e. age, sex, gender, etc.
(iii) Electroencephalographic activity – these will be collected with the minimality invasive and readily available Emotiv EPOC EEG system.
Then i-iii, along with the predetermined ‘audio themes’ will be subject to a, still to be envisaged, process of dimensionality reduction and parametrization which will populate a database.
This database will be subjected to Data Mining proccess in order to extract Mental State Biased Compositional Rules. For instance, decision trees, cluster representaions, or just knowledege from sensitivity analysis.
Pls – feel free to comment by posting in this space.