@MASTERSTHESIS{pgi2020016, author = "A. Giordano", supervisor = "K. Liaskos", title = "Automated personalised music reproducer according to running speed by means of Machine Learning", school = "Department of Computer and Information Sciences, University of Strathclyde", year = "2019", abstract = "In this dissertation, Machine Learning and Neural Network techniques are applied to step recognition and beat detection problem, providing a profound analysis of the full process from the acquisition of the data, to the processing, ending with building and evaluation of the produced models. It has been particularly examined the potentiality of Long Short Term Memory net works applied to both these context. The goal of this dissertation is to prove how it is possible to build an Android's application that by making use of the mentioned techniques could reproduce music with the same tempo as the running pace. After conducting several experiments and testings, the research showed how intriguing results could be achieved on step detection, while further analysis will be needed to beat detection. It has been finally shown how it is achievable to lively match songs and user's pace in an Android's prototype application, which, in future works, would be possible to combine with the designed and proposed neural network models.", }