一种面向机器学习的情境识别机制

A Context Recognition Method Using Machine Learning Algorithm

普适计算环境为了能够自适应地为人们提供透明式的服务,必须通过情境获取和识别。情境识别指通过传感器网络获取物理环境和用户状态等信息后,通过数据融合、规则推理、机器学习等途径,判断当前发生的事务。本文开发了一个情境识别引擎,详细介绍了识别引擎的总体架构和训练算法,并进行了性能检测。

Context recognition is crucial to the pervasive computing environment in order to provide transparent services. It aims at the judgment of the current state of the environment by different methods such as data fusion, logic reasoning and machine learning after the acquisition of context information through sensor network. This paper develops a context recognition engine and discusses the architecture and the training algorithm. The performance evaluation proves the recognition engine is applicable.