Abstract-The location tracking functionality of modern mobile devices provides unprecedented opportunity to the understanding of individual mobility in daily life. Instead of studying raw geographic coordinates, we are interested in understanding human mobility patterns based on sequences of place visits which encode, at a coarse resolution, most daily activities. This paper presents a study on place characterization in people's everyday life based on data recorded continuously by smartphones. First, we study human mobility from sequences of place visits, including visiting patterns on different place categories. Second, we address the problem of automatic place labeling from smartphone data without using any geo-location information. Our study on a large-scale data collected from 114 smartphone users over 18 months confirms many intuitions, and also reveals findings regarding both regularly and novelty trends in visiting patterns. Considering the problem of place labeling with 10 place categories, we show that frequently visited places can be recognized reliably (over 80%) while it is much more challenging to recognize infrequent places.

This paper presents an overview of the Mobile Data Challenge (MDC), a large-scale research initiative aimed at generating innovations around smartphone-based research, as well as community-based evaluation of mobile data analysis methodologies. First, we review the Lausanne Data Collection Campaign (LDCC) -an initiative to collect unique, longitudinal smartphone data set for the MDC. Then, we introduce the Open and Dedicated Tracks of the MDC; describe the specific data sets used in each of them; discuss the key design and implementation aspects introduced in order to generate privacypreserving and scientifically relevant mobile data resources for wider use by the research community; and summarize the main research trends found among the 100+ challenge submissions. We finalize by discussing the main lessons learned from the participation of several hundred researchers worldwide in the MDC Tracks.

Human behavior is often complex and context-dependent. This paper presents a general technique to exploit this "multidimensional" contextual variable for human mobility prediction. We use an ensemble method, in which we extract different mobility patterns with multiple models and then combine these models under a probabilistic framework. The key idea lies in the assumption that human mobility can be explained by several mobility patterns that depend on a subset of the contextual variables and these can be learned by a simple model. We showed how this idea can be applied to two specific online prediction tasks: what is the next place a user will visit? and how long will he stay in the current place?. Using smartphone data collected from 153 users during 17 months, we show the potential of our method in predicting human mobility in real life.

No abstract

Human mobility prediction is an important problem which has a large number of applications, especially in context-aware services. This paper presents a study on location prediction using smartphone data, in which we address modeling and application aspects. Building personalized location prediction models from smartphone data remains a technical challenge due to data sparsity, which comes from the complexity of human behavior and the typically limited amount of data available for individual users. To address this problem, we propose an approach based on kernel density estimation, a popular smoothing technique for sparse data. Our approach contributes to existing work in two ways. First, our proposed model can estimate the probability that a user will be at a given location at a specific time in the future, by using both spatial and temporal information via multiple kernel functions. Second, we also show how our probabilistic framework extends to a more practical task of location prediction for a time window in the future. Our approach is validated on an everyday life location datasets consisting of 133 smartphone users. Our method reaches an accuracy of 84% for the next hour, and an accuracy of 77% for the next three hours.