Using Machine Learning for In-Out decision accuracy for venue owner definable services

Abstract

Presence confirmation for being inside certain venue becomes matter of more importance when venue owner might have option to restrict or to provide value added contents for the user per its presence in a given venue during a given time window. In this paper, machine learning is applied to find the confidence of decision about a User Equipment (UE) presence inside a designated venue based on the accumulated data set used for learning. 20 UEs are used such that some are placed inside venue and other outside to collect data set to be used for ML algorithms. The outside locations are the possible human movement areas around the venue. The UEs works as reference data collection sources both from outside and inside. The received mobile network info by each UE is collected over extended time. Data is labeled based on the actual positions of the UEs. Using Python, Machine Learning is applied with very encouraging results to conclude the presence confirmation inside venue or the other way around. Hyper parameter tuning is applied for kNN ML algorithm.