一种基于机器学习的PM2.5精密测量系统设计方法

Design of fine measurement system for PM2.5 based on machine learning

普通的空气质量传感器由于设备本身的测量准确度不高,往往不能很好地反映空气质量的真实情况,特别是比较难以准确获得PM2.5指数。本文提出了一种利用多台低成本传感装置组成测量系统,通过对历史检测得到的大量数据进行机器学习,来提高PM2.5测量准确度的设计方法。通过验证,该测量系统与单个传感装置相比可以提升PM2.5的测量准确度至少15%以上。

Due to low accuracy of common air quality sensors, the real situation of the air quality especially in terms of PM2.5 index cannot be attained easily. In the paper, a measuring system made up of several common air quality sensors is proposed. According to historical big data, the measuring accuracy is improved through machine learning in the design. After verification, the measuring accuracy by the system could be at least15% higher than one single common sensor.