Abstract | ||
---|---|---|
Various activity-based services have been created for use by smartphone users. In the field of activity recognition, researchers frequently use smartphones or devices equipped with built-in sensors to estimate activities. However, in contrast to wristwatch devices that are worn on the arm, users may change the position of the smartphone depending on their situation; this may include placing the device in a bag or pocket. Therefore, a change in the device position should be considered when estimating activities using a smartphone. Considerable research has been conducted under conditions in which a smart phone is placed in a trouser pocket, however, few studies have focused on the changing context and location of the smartphone. Using the Support Vector Machine (SVM) on an Android smartphone, this paper classifies seven types of activity with three types of smartphone position. The results of an experiment conducted with seven smartphone users, indicate that seven possible states were classified with an average accuracy of greater than 95.75%, regardless of the device position. |
Year | DOI | Venue |
---|---|---|
2017 | 10.1007/978-3-319-65521-5_104 | ADVANCES IN NETWORK-BASED INFORMATION SYSTEMS, NBIS-2017 |
Keywords | Field | DocType |
Smartphone,Activity recognition,Device position,Accelerometer,Barometer | Android (operating system),Activity recognition,Computer science,Accelerometer,Support vector machine,Computer network,Human–computer interaction,Barometer | Conference |
Volume | ISSN | Citations |
7 | 2367-4512 | 0 |
PageRank | References | Authors |
0.34 | 7 | 3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Yuki Oguri | 1 | 2 | 0.73 |
Shogo Matsuno | 2 | 0 | 1.01 |
Minoru Ohyama | 3 | 33 | 9.48 |