The task of rain detection, or wet-dry classification
using measurements from commercial microwave links (CMLs)
is a subject that been studied in depth. However, these studies
are based on direct measurement of the signal level, which
is known to be attenuated by rain. In this paper we present,
for the first time an empirical study on rain classification using
records of transmissions errors in the CMLs. Based on a dataset
of measurements taken from operational cellular backhaul
networks and meteorological measurements, and using long
short-term memory (LSTM) units with a multi-variable time
series, we demonstrate that measurements of microwave link
error are related to rain and can even be used for rain detection
(wet-dry classification). We evaluate the performance of LSTM
on CMLs empirically, and analyze the results by comparison
with rain detection based on attenuation measurements in the
same links