Abstract
We propose a three-way dynamic threshold panel model defined without time-varying covariates. The model is thought to detect changes in fruit ripeness observed at several frequencies of electrical impedance spectroscopy. The transition variable is the lagged dependent variable, which is fruit bioimpedance, and the threshold parameter varies by the level of the third way, which is the electrical frequency of the spectroscope. Consistently with the applied aim, we assume cross-independence in the third way. We apply the model to an innovative dataset containing bioimpedance measurements on on-plant strawberries. Our findings suggest that there is a physio-chemical change in the observed strawberries during days 6-7 of the maturation process irrespective of the electrical frequency.