Automating the interpretation of PM _2.5 time‐resolved measurements using a data‐driven approach

Abstract

The rapid development of automated measurement equipment enables researchers to collect greater quantities of time-resolved data from indoor and outdoor environments. While significant, the interpretation of the resulting data can be a timeconsuming effort. This paper introduces an automated process of interpreting PM 2.5 time-resolved data and differentiating PM 2.5 emissions resulting from indoor and outdoor sources. We use Random Forest (RF), a machine learning approach, to study a dataset of 836 indoor emission events that occurred over a 2-week period in 18 apartments in California. In this paper, we show model development and evaluate its performance as the sample size and source vary. We discuss the characteristics of the dataset that tended to help the source identification and why. For example, we show that data from many events and from different apartments are essential for the model to be suitable for analyzing a new separate dataset. We also show that longitudinal data appear to be more helpful than the time frequency of measurements within a given apartment. We use the resulting RF model to analyze PM 2.5 data of an entirely separate dataset collected from 65 new homes in California. The RF model identifies 442 indoor emission events, with only a few misidentifications.

Simulation-Based Analysis of Impacts of Reduced Envelope and Duct Air Leakage on Indoor Air Pollutant Concentrations in Occupied Manufactured Homes

“…2.2.3.3 PM 2.5 Emissions from Occupants' Daily Activities Continuous PM 2.5 emissions from human daily activity were calculated using a data set consisting of measurements from 70 single-family houses and 23 low-income apartments built from 2011 to 2017 in California. In this dataset, the indoor PM emission events were first identified by applying a machine learning approach to the time-resolved PM concentrations measured in each home (Tang et al, 2021;Zhao et al, 2020). That resulted in 5114 mg total PM mass emitted from 53 measured single family houses and 2825 mg total PM mass emitted from 20 low-income apartments.…”

Effective Kitchen Ventilation for Healthy Zero Net Energy Homes with Natural Gas

“…PM2.5 emissions were identified by applying a machine learning approach called Random Forest (RF) to the time-resolved PM2.5 concentration measured in the living room of each home, as described in detail elsewhere (Tang, Chan, and Sohn 2020). Briefly, the RF model was originally developed using a training dataset where the indoor and outdoor PM2.5 concentrations were collected from 18 California low-income apartments (W. R. Chan et al 2018).…”

Factors Impacting Range Hood Use in California Houses and Low-Income Apartments

“…PM 2.5 emissions were identified by applying a machine learning approach called Random Forest (RF) to the time-resolved PM 2.5 concentration measured in the living room of each home, as described in detail elsewhere [48]. Briefly, the RF model was originally developed using a training dataset where the indoor and outdoor PM 2.5 concentrations were collected from 18 California low-income apartments [49].…”