A markov model for Temporal Dominance of Sensations (TDS) data
TDS data can be characterized by a series of TDS dyads that represent the selection of a single dominant attribute, which is updated over time. As such, TDS data describes the presence of different dominant attributes in food products. We analyze consumer TDS evaluations of six flavoured fresh cheeses (Schlich & Thomas, 2014) using a time homogeneous Markov chain. Our fitted models assume that the TDS dyads adhere to the Markovian property and quantifies the probabilities of transitioning between the eight dominant attributes used by consumers to characterize the samples. The probabilities are summarized in transition matrices, which we use to compare the products. Our results show a natural progression of the attributes which can be explained from both physiological and sensory perspectives.
Franczak, B. C., Browne, R. P., Castura, J., Findlay, J., & McNicholas, P. D. (2015). A markov model for Temporal Dominance of Sensations (TDS) data. In: 11th Pangborn Sensory Science Symposium. 23-27 August. Gothenburg, Sweden.