Radio spectrum is a finite resource, while the demand for wireless systems is increasingrnat an exponential rate. To meet this demand, new generations of cellular networks werernintroduced.Spectrum utilization of cellular bands is analyzed widely using spectrum measurements.rnKnowledge of spectrum utilization will help operators like Ethio telecom tornunderstand and plan band usage.rnIn this thesis, using the K-means algorithm and Deep learning algorithms, namely ConvolutionalrnNeural Network (CNN) and Long Short Term Memory (LSTM), downlink GlobalrnSystem for Mobile Communication (GSM) 900 spectrum utilization is analyzed and modeledrnto know the spectrum utilization of Ethio telecom. The data is collected from AddisrnAbaba 639 GSM base stations. Spectrum utilization is modeled using CNN and LSTMrnalgorithms for clustered and non-clustered data. Because of the differences in base stationrnbehavior, clustering base stations is done and model the spectrum utilization of the basernstations in each cluster.rnOur results show that the GSM 900 downlink spectrum is not utilized optimally. Thernhighest observed average spectrum utilization was 71%, with the lowest observed averagernspectrum utilization being 1.4%. The model developed for the cluster data using the CNNrnalgorithm can model spectrum utilization with an RMSE value of 0.58 and this model canrnpredict the next twenty-four-hour base station spectrum utilization with an RMSE valuernof 1.04.