스마트폰 사용자의 효율적인 개인 분류를 위한 특징 선택
Feature Selection for Efficient Person Classification of Smartphone Users
0
조회
0
다운로드
1
인용
0
참고문헌
키워드
초록
최근 사용자 움직임을 감지할 수 있는 센서가 스마트폰에 내장되면서 이를 활용하여 사용자들이 편리하게 스마트폰을 사용하게 하려는 다양한 연구가 이루어지고 있으나, 이러한 지속적인 처리에 드는 계산 비용과 배터리 소모 및 저장 공간 문제로 어려움을 겪고 있다. 특히 사용자 분류를 위한 기계학습 과정에서 중요한 최적의 특징을 선택하여 사용하는 방안에 대한 연구가 부족한 실정이다. 본 논문에서는 효율적으로 사용자를 분류하기 위한 특징 선택을 제안한다. 이를 위하여 공개 데이터인 UCI HAR 데이터 세트의 특징을 사용하였으며, 실험은 오픈소스 소프트웨어인 WEKA를 사용하였다. 시간 영역 특징 중에서 최적의 특징 세트를 선정하기 위하여 먼저 다섯 가지 특징 선택 방법으로 특징을 선택하였다. 다음으로 선택된 특징들에서 특징의 중복 정도로 구분하여 세 가지 특징 부분 세트를 생성하였다. 이 세트를 입력으로 여섯 가지 기계 학습 분류 기법을 가지고 실험하였으며, 그 결과로 전체 특징 중 약 5.9%를 사용한 특징 부분 세트에서 전체 특징 세트를 사용한 경우보다 개선된 93.6%의 정확도를 구하였다.
Abstract
Recently, as sensors that can detect user movements have been built into the smartphone, various research are being conducted to make users conveniently use the smartphone by utilizing them. However, it suffers from computational cost, battery drain, and storage space issues for such continuous processing. In this paper, we propose a feature selection for efficient user classification. In particular, there is a lack of research on how to select and use the optimal features in the machine learning process for user classification. In this paper, we propose a feature selection for efficient user classification. For this, the features of the open HAR dataset in the UCI Machine Learning Repository were used, and experiments were conducted using the open-source software WEKA. In order to select an optimal feature set from the time domain features, five feature subsets were first generated using the five feature selection methods. Next, the three feature subsets were generated by dividing the degree of overlapping of the features in the five previously selected subsets. These sets were tested with the six machine learning classification techniques as inputs. As a result, the accuracy of 93.6%, which is improved than the case of the using the entire features set, was obtained from the feature subset using about 5.9% of the entire features.
참고문헌 (22)
- J-H. Kim, "Personal authentication technology and database construction using biosignals", Journal of Telecommunication Technology Association, 165, pp.41~46 (2016)
- M. EhatishamulHaq, "Authentication of smartphone users based on activity recognition and mobile sensing", Sensors, 17(9), pp.2043~ (2017)
- K. Chen, "Deep learning for sensor-based human activity recognition: overview, challenges and opportunities", arXiv preprint arXiv:2001.07416, ~ (2020)
- E. Pippa, "Feature selection evaluation for light human motion identification in frailty monitoring system", ICT4Ageing, pp.88~95 (2016)
- J. C. Quiroz, "Feature selection for activity recognition from smartphone accelerometer data", Intelligent Automation & Soft Computing, pp.1~9 (2017)
- W. W. Myo, "A Cyclic attribution technique feature selection method for human activity recognition", IJ Intell. Syst. Appl, 10, pp.25~32 (2019)
- N. Ahmed, "Enhanced human activity recognition based on smartphone sensor data using hybrid feature selection model", Sensors, 20(1), pp.317~ (2020)
- C. Tang, "An empirical evaluation of activities and classifiers for user identification on smartphones", IEEE 8th International Conference on Biometrics Theory, Applications and Systems, pp.1~8 (2016)
- 김영인, "스마트폰 센서를 이용한 사용자 행위 기반 개인 분류 기법의 성능 평가", 한국지식정보기술학회 논문지, 13(6), pp.835~845 (2018), 한국지식정보기술학회[DOI]
- D. Figo, "Preprocesing techniques for context recognition from acelerometer data", Personal and Ubiquitous Computing, 14(7), pp.645~662 (2010)
- D. Anguita, "A public domain dataset for human activity recognition using smartphones", European Symposium on Artificial Neural Networks, pp.437~442 (2013)
- I. Kononenko, "Estimating attributes:Analysis and extension of RELIEF", Proc. Euro. Conf. Mach. Learn, 784, pp.171~182 (1994)
- J-L. Reyes-Ortiz, "Transition-aware human activity recognition using smartphones", Neurocomputing, 171, pp.754~767 (2016)
- M. EhatishamulHaq, "Continuous authentication of smartphone users based on activity pattern recognition using passive mobile sensing", Journal of Network and Computer Applications, 109, pp.24~35 (2018)
- E. Frank, "Online appendix for data mining: Practical machine learning tools and techniques" (2016), Morgan Kaufmann
- D. P. Ismi, "K-means clustering based filter feature selection on high dimensional data", International Journal of Advances in Intelligent Informatics, 2(1), pp.38~45 (2016)
- "Package weka.attributeSelection" (2021)[URL]
- M. A. Hall, "Correlation-based feature subset selection for machine learning", Ph. D, pp.1998
- H. Wang, "Ensemble feature selection technique for software quality classification", Proceedings of the 22nd International Conference on Software Engineering & Knowledge Engineering, pp.215~220 (2010)
- R. Kohavi, "A study of cross-validation and bootstrap for accuracy estimation and model selection", International Joint Conferences on Artificial Intelligence, 14(2), pp.1137~1145 (1995)
- M. Sokolova, "A systematic analysis of performance measures for classification tasks", Information processing & management, 45(4), pp.427~437 (2009)
- M. Sébastien, "Handbook of biometric anti-spoofing:trusted biometrics under spoofing attacks" (2014), Springer