TY - GEN
T1 - Specifying latent factors with a domain model for personal data analysis
AU - Tago, Kiichi
AU - Takagi, Kosuke
AU - Ito, Kenichi
AU - Jin, Qun
N1 - Funding Information:
The work has been partially supported by 2016–2018 Masaru Ibuka Foundation Research Project on Oriental Medicine.
Publisher Copyright:
© 2018 IEEE.
PY - 2018/10/26
Y1 - 2018/10/26
N2 - Personal data is data related to an individual, generated by an individual, or metadata about an individual. To analyze personal data comprehensively, it is needed to consider different types and sources of data. Moreover, it should be considered not only explicit attributes but also latent factors. In this study, to specify latent factors, we use Structural Equation Modeling (SEM) with a domain model for personal data analysis. The domain model represents the relationship between the latent factors and measures that are possible to be obtained by a wearable device. We construct an activeness model as the domain model and apply it for personal data analysis. The activeness level which is assumed as the latent factor is quantified by SEM. We verify the adaptability of the activeness model by comparing the case of classifying by the activeness factor with the case of not using latent factors. The result shows that the model has higher adaptability when personal data is classified by latent factors than only by labels.
AB - Personal data is data related to an individual, generated by an individual, or metadata about an individual. To analyze personal data comprehensively, it is needed to consider different types and sources of data. Moreover, it should be considered not only explicit attributes but also latent factors. In this study, to specify latent factors, we use Structural Equation Modeling (SEM) with a domain model for personal data analysis. The domain model represents the relationship between the latent factors and measures that are possible to be obtained by a wearable device. We construct an activeness model as the domain model and apply it for personal data analysis. The activeness level which is assumed as the latent factor is quantified by SEM. We verify the adaptability of the activeness model by comparing the case of classifying by the activeness factor with the case of not using latent factors. The result shows that the model has higher adaptability when personal data is classified by latent factors than only by labels.
KW - Latent factors
KW - Personal data analysis
KW - Structural Equation Modeling (SEM)
UR - http://www.scopus.com/inward/record.url?scp=85056863184&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85056863184&partnerID=8YFLogxK
U2 - 10.1109/DASC/PiCom/DataCom/CyberSciTec.2018.00056
DO - 10.1109/DASC/PiCom/DataCom/CyberSciTec.2018.00056
M3 - Conference contribution
AN - SCOPUS:85056863184
T3 - Proceedings - IEEE 16th International Conference on Dependable, Autonomic and Secure Computing, IEEE 16th International Conference on Pervasive Intelligence and Computing, IEEE 4th International Conference on Big Data Intelligence and Computing and IEEE 3rd Cyber Science and Technology Congress, DASC-PICom-DataCom-CyberSciTec 2018
SP - 292
EP - 299
BT - Proceedings - IEEE 16th International Conference on Dependable, Autonomic and Secure Computing, IEEE 16th International Conference on Pervasive Intelligence and Computing, IEEE 4th International Conference on Big Data Intelligence and Computing and IEEE 3rd Cyber Science and Technology Congress, DASC-PICom-DataCom-CyberSciTec 2018
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 16th IEEE International Conference on Dependable, Autonomic and Secure Computing, IEEE 16th International Conference on Pervasive Intelligence and Computing, IEEE 4th International Conference on Big Data Intelligence and Computing and IEEE 3rd Cyber Science and Technology Congress, DASC-PICom-DataCom-CyberSciTec 2018
Y2 - 12 August 2018 through 15 August 2018
ER -