摘要
在居家养老服务中,如何获取老人居家活动,是精确主动为老人提供居家服务的关键问题。本文以居家养老服务为研究背景,首先分析了居家养老服务中活动的类别,以及各类活动所包含的行为。然后建立了老人居家行为识别问题模型,阐述了使用情境感知技术来获取老人行为的解决思路,进而提出了基于随机森林和行为相似性的两层行为识别算法。最后通过大量实验验证了算法的正确性与性能。本文提出的行为识别算法能够准确识别老人行为,其准确率可达到95.59%,效果优于同类其它方法。
In home-based care service,how to precisely obtain behaviors of the elderly is a key issue to provide high-quality home services for them.Thus,this paper studies the types of activities in home-based care service and the relationship between activities and behaviors.Afterwards,the paper proposes a model of behavior recognition for the elderly at home and decides to use context-aware techniques to obtain the behavior of the elderly.Then,a two-layer behavior recognition algorithm based on Random Forest and behavioral similarity is proposed.Finally,the paper conducts a series of experiments to show the correctness and performance of the algorithm.The proposed algorithm for elderly people in this paper can accurately identify the behavior,and the accuracy rate is over 95.59%,which is better than other methods.
作者
潘宇欣
郑彬
张龙
于鹏飞
徐汉川
PAN Yuxin;ZHENG Bin;ZHANG Long;YU Pengfei;XU Hanchuan(School of Computer Science and Technology,Harbin Institute of Technology,Harbin 150001,China;Shangdong University,Jinan 250100,China;Shandong Yidong Network Technology Co.,Ltd.,Jinan 250100,China)
出处
《智能计算机与应用》
2019年第5期312-319,共8页
Intelligent Computer and Applications
基金
国家重点研发计划资助项目(2018YFB1402901)
关键词
居家养老服务
情境感知
行为识别
随机森林
行为相似性
home-based care service
context awareness
behavior recognition
Random Forest
behavioral similarity
