Scholarlink

This chapter explores the crucial steps of data preprocessing in air quality monitoring, focusing on missing value imputation and outlier detection. For missing value imputation, both univariate and multivariate methods are introduced, with examples of the former including linear interpolation, EM algorithms, and neural networks, and the latter including self-organizing maps (SOM) and multivariate nearest neighbor methods. Regarding outlier detection, the chapter discusses three approaches: unsupervised, filtering, and forecasting. Air quality data from the Jing-Jin-Ji region is utilized to analyze the temporal and spatial characteristics of pollutants, supporting the proposed data preprocessing methods. The analysis reveals that pollutant concentrations exhibit clear periodicity and spatial correlation.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Data Preprocessing in Air Quality Monitoring

  • 章节

  • 发表日期：

    2025年

查看全文
  • Hui Liu,
  • Yanfei Li,
  • Zhu Duan

摘要

This chapter explores the crucial steps of data preprocessing in air quality monitoring, focusing on missing value imputation and outlier detection. For missing value imputation, both univariate and multivariate methods are introduced, with examples of the former including linear interpolation, EM algorithms, and neural networks, and the latter including self-organizing maps (SOM) and multivariate nearest neighbor methods. Regarding outlier detection, the chapter discusses three approaches: unsupervised, filtering, and forecasting. Air quality data from the Jing-Jin-Ji region is utilized to analyze the temporal and spatial characteristics of pollutants, supporting the proposed data preprocessing methods. The analysis reveals that pollutant concentrations exhibit clear periodicity and spatial correlation.