Data collection from mobile phone for personalized behaviour mining (transportation mode)

Abstract

As humans share an ever increasing amount of location information online through

location enabled social networks, an increasing amount of people are looking to

stay in control of their own information. The most common method of data

collection of oneself is the mobile phone with its ever increasing number of sensors.

This report is about the usage of a mobile phone to collect information; specifically

location information in order to build personalized models of an individual’s

movement patterns and habits. Other information like public transport route data is

also used to supplement the models. The models are then used to predict the

individual’s location.

Three models are built, namely: are a spatial model which had an accuracy of 25%, a

spatial-temporal model with an accuracy of 29%, and a public transport analysis

model which had an accuracy of 98% in finding possible transport service along a

segment of the route 3 stops long.