The proposed invention is a method and system for the segmentation of dual-axis accelerometry signals for the purposed of identifying problematic swallowing events. The method and system employ a sensor, a data collection means including an algorithm for the analysis of the data. The proposed invention considers the stochastic properties of swallowing signals in both directions, A-P and S-I to extract events associated with swallowing. A segmentation algorithm may be applied to the signals to establish the time duration of swallows and swallows may be classified with respect to gender, body mass index, age or types of swallow.