This paper investigates the application of machine learning for optimal deployment of 5G infrastructure, such as the position and the orientation of the antenna that help achieve the best signal coverage. This avoids the need to perform on-site measurements or extensive software simulations. Multivariate Regression (MR) and Neural Network (NN) models are applied to predict the signal coverage in an indoor environment. The results show that the average prediction error using NN for the case investigated is 7 dB for a 60-GHz operating frequency, whereas the error using the MR technique is lower than 6 dB. The unique aspect in the investigation is the integration of the clustering algorithm and the NN machine learning model for predicting indoor signal coverage.
