基于mESM的情境主观数据收集实验研究  

THE RESEARCH BASED ON MESM FOR SUBJECTIVE CONTEXT DATA COLLECTION

作　　者：刘迪 刘正捷[1] 刘伟 Liu Di;Liu Zhengjie;Liu Wei(School of Information Science and Technology,Dalian Maritime University,Dalian 116026,Liaoning,China)

机构地区：[1]大连海事大学信息科学技术学院,辽宁大连116026

出　　处：《计算机应用与软件》2023年第11期100-106,162,共8页Computer Applications and Software

摘　　要：移动体验抽样方法(mobile Experience Sampling Method,mESM)已被广泛用于即时临场的移动健康数据收集,从而克服了临床评估的回忆偏倚问题。尽管使用mESM在许多研究中具有很多优势,但是使用mESM获得高质量数据仍然是一个挑战。通过智能手机传感器数据提取出情境特征,探讨广泛情境因素对参与者响应率的影响,实现和评估两阶段机器学习模型,根据情境特征预测mESM触发时的参与者的响应率和响应延迟。通过实例研究,研究结果突出了一些对参与者的响应率相关的情境因素。旨在设计一个智能mESM系统,以提升参与者的参与度并提高数据收集的质量。Mobile device-based experience sampling method(mobile ESM,mESM)has been widely used to collect instant and in-situ mobile health data,which overcomes recall bias of in-clinic assessment.It has many advantages using mobile experience sampling method(mESM)in many studies,but collecting high quality data with mESM is challenging.This paper extracted contextual features via smartphone sensor to explore the effect of different contextual factor on participants'response rate.This paper implemented and evaluated a two-stage hierarchical machine-learning model to predict participants'response rate and response delay for mESM delivery based on contextual features.Through the case study,experimental results highlight a number of factors correlated with participants'response rate.This work aims at designing an intelligent mESM system that increases participant engagement and improves the quality of data collection.

关 键 词：移动体验抽样方法 情境感知 移动计算 移动感知 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]