基于时间序列分析的环境γ辐射剂量率数据预处理方法研究及评估

Research and evaluation of natural environmentalγradiation dose rate data preprocessing method based on time sequences analysis

环境γ辐射剂量率数据处理与利用已经成为目前环境质量监测领域的热点之一。本文通过对γ辐射剂量率数据进行统计学分析、清洗及降噪等处理,提出了基于时间序列分析的γ辐射剂量率数据预处理方法,设置了基于长短期记忆网络(Long Short-Term Memory,LSTM)的有监督特殊数据检测模型,并评估包含数据集成、数据分析、数据清洗、数据变换、数据转换的数据预处理方法对特殊数据检测模型的影响。结果表明,经数据预处理后,数据质量提高,特殊数据识别在准确率、精确率、召回率、F1-分数方面得到了明显改善,数据预处理为后续进一步的数据挖掘及特殊数据研究奠定了良好基础。

In recent years,data preprocessing and utilization of natural environmentalγradiation dose rate have become one of the hot spots in the field of environmental quality monitoring.This paper proposes data preprocessing procedures through statistical data analysis,data cleaning,and time sequences denoising.And this paper also investigates the impact of data preprocessing with evaluation to provide high-quality data for outliers detection techniques based on long short-term memory(LSTM).After data preprocessing,data quality is improved,and the outliers detection is significantly improved in terms of accuracy,precision,recall,and F1-score ect.Data preprocessing lays a good foundation for further data mining and research of outliers.