一种基于事件的Lifelog管理模型

A Lifelog Management Model Based on Events

摘要: 从2011年开始，我们发起了一个LiuLifelog项目，该项目收集了22位志愿者的3万条lifelogging数据。但是随着时间的推移，用户的lifelogging数量越来越多，管理这些数据对于用户来说变得越发复杂。为此，本文提出了一种LE-PDA模型，该模型将lifelogging数据划分为不同的事件，同时为每一个事件添加了标签，方便用户管理数据。在本次实验中，两位用户的数据在应用LE-PDA模型后，分别被划分为5215个和1086个带有标签的事件，且准确率达到了68%。实验证明了LE-PDA模型在LiuLifelog数据集上的实用性，它能帮助用户更高效的组织和管理lifelogging数据。

Abstract: Since 2011, we have initiated the LiuLifelog project, which has collected 30,000 lifelogging data en-tries from 22 volunteers. However, as time progresses, the quantity of users’ lifelogging data in-creases, making the management of these data increasingly complex for the users. To address this, this paper proposes a LE-PDA model, which segments lifelogging data into distinct events, while as-signing labels to each event, facilitating user data management. In this experiment, the data from two users were divided into 5,215 and 1,086 labeled events respectively, after applying the LE-PDA model, achieving an accuracy rate of 68%. The experiment validates the efficacy of the LE-PDA model on the LiuLifelog dataset, demonstrating its ability to assist users in organizing and managing lifelogging data more efficiently.