Name
Abstract | ||
|---|---|---|
This paper presents our recent work on the analyses of smart phone sensor data collected for the human activity recognition (HAR), with the objective to develop more accurate activity recognition systems independent of smart phone models. We identify the multi-device scenario and present the impairments of different smartphone embedded sensor models on HAR applications. Outlier removal, interpolation, and filters in the preprocessing stage are proposed as mitigating techniques. Based on datasets collected from four distinct smartphones, the proposed mitigating methods show positive effects on 10-fold cross validation, device-to-device validation, and leave-one-out validation. Improved performance for smartphone based human activity recognition is observed. |
Year | Venue | Keywords |
|---|---|---|
2016 | 2016 IEEE INTERNATIONAL CONFERENCE ON SYSTEMS, MAN, AND CYBERNETICS (SMC) | Human activity recognition, smart phone sensors, filter, interpolation, outliers, machine learning |
Field | DocType | ISSN |
Activity recognition,Accelerometer,Computer science,Interpolation,Software,Preprocessor,Phone,Artificial intelligence,Smart phone,Cross-validation,Machine learning | Conference | 1062-922X |
Citations | PageRank | References |
0 | 0.34 | 0 |
Authors | ||
4 | ||
Name | Order | Citations | PageRank |
|---|---|---|---|
| Xizhe Yin | 1 | 3 | 1.06 |
| Gary Shen | 2 | 0 | 0.34 |
| Xianbin Wang | 3 | 2365 | 223.86 |
| Weiming Shen | 4 | 3407 | 343.73 |