Name
Abstract | ||
|---|---|---|
Commuting matrices are key for a variety of fields, including transportation engineering and urban planning. Up to now, these matrices have been typically generated from data obtained from surveys. Nevertheless, such approaches typically involve high costs which limits the frequency of the studies. Cell phones can be considered one of the main sensors of human behavior due to its ubiquity, and as a such, a pervasive source of mobility information at a large scale. In this paper we propose a new technique for the estimation of commuting matrices using the data collected from the pervasive infrastructure of a cell phone network. Our goal is to show that we can construct cell-phone generated matrices that capture the same patterns as traditional commuting matrices. In order to do so we use optimization techniques in combination with a variation of Temporal Association Rules. Our validation results show that it is possible to construct commuting matrices from call detail records with a high degree of accuracy, and as a result our technique is a cost-effective solution to complement traditional approaches. |
Year | DOI | Venue |
|---|---|---|
2012 | 10.1145/2346496.2346499 | UrbComp@KDD |
Keywords | Field | DocType |
traditional approach,high degree,optimization technique,cellphone network data,cell phone network,commuting matrix,pervasive infrastructure,cell phone,high cost,pervasive source,new technique,urban planning,data collection,human behavior,cost effectiveness,association rule | Data mining,Commuting matrices,Computer science,Matrix (mathematics),Association rule learning,Urban planning,Network data,Cellular network | Conference |
Citations | PageRank | References |
21 | 1.13 | 7 |
Authors | ||
3 | ||
Name | Order | Citations | PageRank |
|---|---|---|---|
| Vanessa Frias-Martinez | 1 | 213 | 17.79 |
| Cristina Soguero | 2 | 21 | 1.13 |
| Enrique Frias-Martinez | 3 | 238 | 17.11 |