The proliferation of mobile devices has led to an increasing demand for radio spectrum resources. Currently, the spectrum allocation is static, which has resulted in underutilization of this resource. This situation has motivated the search for solutions to address the spectrum scarcity problem. The channel occupancy rate is a piece of information that can assist the decision making process regarding reallocation of spectral resources. This information can be useful for DSA (Dynamic Spectrum Access) systems, spectrum decision, and spectrum regulators. In most of the spectrum usage surveys, frequentist inference techniques are used to scan the spectrum and estimate the occupancy rate of particular frequency bands. This approach has several limitations that are addressed in this work through Bayesian inference in order to build a spectrum decision framework able to deal with uncertainty. This paper presents a work in progress, whose goal is to build a probabilistic model to assist the spectrum decision process in cognitive radios.