基于加权朴素贝叶斯的水质数据分类研究

doi:10.11959/j.issn.2096-3750.2022.00255

物联网学报 ›› 2022, Vol. 6 ›› Issue (1): 113-122.doi: 10.11959/j.issn.2096-3750.2022.00255

基于加权朴素贝叶斯的水质数据分类研究

方志豪¹, 李正权¹^,², 张铭玮¹

¹ 江南大学物联网工程学院，江苏无锡 214122
² 江苏省未来网络创新研究院，江苏南京 211111

修回日期:2021-12-29 出版日期:2022-03-30 发布日期:2022-03-01
作者简介:方志豪（1996− ），男，江南大学物联网工程学院硕士生，主要研究方向为水质监测系统应用和开发
李正权（1976− ），男，江南大学物联网工程学院教授，主要研究方向为大规模MIMO技术、协作通信、物联网等
张铭玮（1998− ），男，江南大学物联网工程学院硕士生，主要研究方向为水质监测系统应用和开发
基金资助:
国家自然科学基金资助项目(61571108);无锡市科技发展资金资助项目(H20191001);无锡市科技发展资金资助项目(G20192010);未来网络科研基金项目(FNSRFP-2021-YB-11)

Research on water quality data classification based on weighted Naive Bayes

Zhihao FANG¹, Zhengquan LI¹^,², Mingwei ZHANG¹

¹ School of Internet of Things Engineering, Jiangnan University, Wuxi 214122, China
² HJiangsu Future Networks Innovation Institute, Nanjing 211111, China

Revised:2021-12-29 Online:2022-03-30 Published:2022-03-01
Supported by:
The National Natural Science Foundation of China(61571108);The Wuxi Science and Technology Development Fund(H20191001);The Wuxi Science and Technology Development Fund(G20192010);The Future Network Scientific Research Fund Project(FNSRFP-2021-YB-11)

摘要/Abstract

摘要：

为更好地实施水环境管理政策，水质评价是基础环节，即根据某一水域多个水质参数，如何将其合理地划分到特定水质类别。针对该问题，提出了一种改进的朴素贝叶斯分类方法，该方法赋予不同属性以不同的权值，削弱了朴素贝叶斯条件独立性的假设，使分类结果更接近实际类别。首先，参考国家地表水水质自动监测站（以下简称国控水站）发布的数据，选取其中500条水质数据作为样本，基于溶解氧、高锰酸盐指数、氨氮和总磷4个指标建立评价体系；然后，利用改进朴素贝叶斯分类方法对样本进行学习与评价，并采用五折交叉验证法验证其分类性能。结果表明，改进朴素贝叶斯分类方法的准确率、精确率、召回率和F1值分别达到96.0%、95.9%、93.8%和94.8%，水质数据分类的性能指标相较于其他朴素贝叶斯分类方法更高，可对实际工程中遇到水质数据分类的问题提供一定的参考。

关键词: 水质评价, 朴素贝叶斯, 五折交叉验证, 性能指标

Abstract:

In order to better implement the water environmental management policies, water quality evaluation is the basic step, that is to reasonably divide it into specific water quality category according to multiple water quality parameters in a certain water area.Aimed at this problem, an improved Naive Bayes classification method was proposed, which endowed different attributes with different weights, weakened the assumption of Naive Bayes conditional independence, and made the classification result closer to the actual category.Firstly, referred to the data released by the national surface water quality automatic monitoring station, 500 water quality data were selected as samples, and an evaluation system with four indicators was established, including dissolved oxygen, permanganate index, ammonia nitrogen and total phosphorus.And then, the improved Naive Bayes classification method was used to learn and evaluate the samples, and its classification performance by the five fold cross validation method was verified.The results show that the accuracy, precision, recall and F1 value of the improved Naive Bayes classification method reach 96.0%, 95.9%, 93.8% and 94.8% respectively, with higher performance index of water quality data classification compared with other Naive Bayes classification method, which can provide some reference for the problem of water quality data classification encountered in actual engineering.

Key words: water quality evaluation, Naive Bayes, ive fold cross validation, performance index

中图分类号:

X824

方志豪, 李正权, 张铭玮. 基于加权朴素贝叶斯的水质数据分类研究[J]. 物联网学报, 2022, 6(1): 113-122.

Zhihao FANG, Zhengquan LI, Mingwei ZHANG. Research on water quality data classification based on weighted Naive Bayes[J]. Chinese Journal on Internet of Things, 2022, 6(1): 113-122.

图/表 12

图1

图2

表1

图3

图4

图5

表2

表3

表4

表5

图6

图7

参考文献 20

[1]	AWOKE A , BEYENE A , KLOOS H ,et al. River water pollution status and water policy scenario in Ethiopia:raising awareness for better implementation in developing countries[J]. Environmental Management, 2016,58(4): 694-706.
[2]	宁阳明, 尹发能, 李香波 . 几种水质评价方法在长江干流中的应用[J]. 西南大学学报(自然科学版), 2020,42(12): 126-133.
	NING Y M , YIN F N , LI X B . Application of several evaluation methods for river water quality in the Yangtze River Mainstream[J]. Journal of Southwest University (Natural Science Edition), 2020,42(12): 126-133.
[3]	乔辉 . 基于物联网技术的水质监控系统[J]. 物联网学报, 2017,1(1): 81-85.
	QIAO H . Water quality monitoring system based on IoT[J]. Chinese Journal on Internet of Things, 2017,1(1): 81-85.
[4]	王竹, 朱士江, 刘扬 ,等. 不同水质评价方法在滦河下游段的比较应用[J]. 节水灌溉, 2019(10): 68-72,77.
	WANG Z , ZHU S J , LIU Y ,et al. Comparative application of different water quality evaluation methods in the downstream section of LuanheRiver[J]. Water Saving Irrigation, 2019(10): 68-72,77.
[5]	杨浩, 张国珍, 杨晓妮 ,等. 基于模糊综合评判法的洮河水环境质量评价[J]. 环境科学与技术, 2016,39(S1): 380-386,392.
	YANG H , ZHANG G Z , YANG X N ,et al. Comprehensive evaluation on water environment quality of the Tao river based on fuzzy comprehensive Method[J]. Environmental Science ＆ Technology, 2016,39(S1): 380-386,392.
[6]	王瑶, 王慧勇, 安丽娟 ,等. 黄壁庄水库水质评价及氮污染成因分析[J]. 水电能源科学, 2020,38(4): 60-63.
	WANG Y , WANG H Y , AN L J ,et al. Water quality evaluation and cause analysis of nitrogen pollution in huangbizhuang Reservoir[J]. Water Resources and Power, 2020,38(4): 60-63.
[7]	WU Z S , WANG X L , CHEN Y W ,et al. Assessing river water quality using water quality index in Lake Taihu Basin,China[J]. Science of the Total Environment, 2018,612: 914-922.
[8]	SUN W , XIA C Y , XU M Y ,et al. Application of modified water quality indices as indicators to assess the spatial and temporal trends of water quality in the Dongjiang River[J]. Ecological Indicators, 2016,66: 306-312.
[9]	HOU W , SUN S H , WANG M Q ,et al. Assessing water quality of five typical reservoirs in lower reaches of Yellow River,China:using a water quality index method[J]. Ecological Indicators, 2016,61: 309-316.
[10]	国家环境保护总局,国家质量监督检验检疫总局. 地表水环境质量标准:GB 3838—2002[S]. 北京:中国环境科学出版社, 2002.
	State Environmental Protection Administration of the People's Republic of China,General Administration of Quality Supervision,Inspection and Quarantine of the People's Republic of China. 地表水环境质量标准:GB 3838—2002[S]. Beijing:China Environment Science Press, 2002.
[11]	张颖, 高倩倩 . 基于随机森林分类算法的巢湖水质评价[J]. 环境工程学报, 2016,10(2): 992-998.
	ZHANG Y , GAO Q Q . Water quality evaluation of Chaohu Lake based on random forest method[J]. Chinese Journal of Environmental Engineering, 2016,10(2): 992-998.
[12]	ZHANG R , LIANG S , OU M H ,et al. Evaluation of water quality for mangrove ecosystem using artificial neural networks[C]// Proceedings of 2018 International Conference on Advanced Mechatronic Systems (ICAMechS). Piscataway:IEEE Press, 2018: 257-261.
[13]	王铁良, 苏芳莉, 孙迪 ,等. 基于模糊BP神经网络的辽河口湿地水质评价[J]. 西北林学院学报, 2020,35(5): 195-200.
	WANG T L , SU F L , SUN D ,et al. Water quality evaluation of Liaoheestuary wetland based on back propagation artificial neural Network[J]. Journal of Northwest Forestry University, 2020,35(5): 195-200.
[14]	万哲慧, 王珅, 冯孙林 ,等. 熵权贝叶斯模型在珊溪水库水环境质量评价的应用[J]. 节水灌溉, 2018(3): 55-57,62.
	WAN Z H , WANG S , FENG S L ,et al. Application of entropy weight Bayes model in Shanxi reservoir drinking water sources water quality assessment[J]. Water Saving Irrigation, 2018(3): 55-57,62.
[15]	HAN J , KAMBER M , PEI J . Data mining:concepts and techniques,3rd Edition[M]. San Francisco: Morgan Kaufmann Publishers, 2012.
[16]	张步良 . 基于分类概率加权的朴素贝叶斯分类方法[J]. 重庆理工大学学报(自然科学), 2012,26(7): 81-83.
	ZHANG B L . Naive Bayesian classifier method based on the classification of the probability-Weighted[J]. Journal of Chongqing University of Technology (Natural Science), 2012,26(7): 81-83.
[17]	XU W Q , JIANG L X , YU L J . An attribute value frequency-based instance weighting filter for naive Bayes[J]. Journal of Experimental and Theoretical Artificial Intelligence, 2019,31(2): 225-236.
[18]	陈誉洋 . 基于特征选择和增强训练的朴素贝叶斯网络钓鱼检测[D]. 合肥:安徽大学, 2020.
	CHEN Y Y . Naive Bayes phishing detection based on feature selection and reinforcement training[D]. Hefei:Anhui University, 2020.
[19]	SUBRAMANIAN R S , PRABHA D . Customer behavior analysis using Naive Bayes with bagging homogeneous feature selection approach[J]. Journal of Ambient Intelligence and Humanized Computing, 2021,12(5): 5105-5116.
[20]	陈曦, 张坤 . 一种基于树增强朴素贝叶斯的分类器学习方法[J]. 电子与信息学报, 2019,41(8): 2001-2008.
	CHEN X , ZHANG K . A classifier learning method based on tree-augmented naive Bayes[J]. Journal of Electronics ＆ Information Technology, 2019,41(8): 2001-2008.

指标	Ⅰ类	Ⅱ类	Ⅲ类	Ⅳ类	Ⅴ类
pH值	6～9	6～9	6～9	6～9	6～9
溶解氧/(mg·L^-1)	≥7.5	≥6	≥5	≥3	≥2
高锰酸盐指数/(mg·L^-1)	≤2	≤4	≤6	≤10	≤15
氨氮/(mg·L^-1)	≤0.15	≤0.5	≤1.0	≤1.5	≤2.0
总磷/(mg·L^-1)	≤0.02	≤0.1	≤0.2	≤0.3	≤0.4

混淆矩阵	预测为正	预测为负
实际为正	TP	FN
实际为负	FP	TN

		预测的类
		类=1	类=2	类=3	类=4	类=5
	类=1	d₁₁	d₁₂	d₁₃	d₁₄	d₁₅
	类=2	d₂₁	d₂₂	d₂₃	d₂₄	d₂₅
实际的类	类=3	d₃₁	d₃₂	d₃₃	d₃₄	d₃₅
	类=4	d₄₁	d₄₂	d₄₃	d₄₄	d₄₅
	类=5	d₅₁	d₅₂	d₅₃	d₅₄	d₅₅

	类别1	类别2	类别3	类别4	类别5
标准朴素贝叶斯分类方法	92.6%	93.5%	95.5%	95.3%	93.9%
文献[16]所提方法	90.9%	94.4%	94.8%	95.5%	96.9%
文献[17]所提方法	96.2%	95.3%	96.3%	95.4%	88.9%
本文所提方法	98.0%	96.2%	95.6%	95.5%	94.4%

	类别1	类别2	类别3	类别4	类别5
标准朴素贝叶斯分类方法	100.0%	99.5%	94.1%	84.7%	77.5%
文献[16]所提方法	100.0%	99.0%	94.1%	87.5%	77.5%
文献[17]所提方法	100.0%	99.5%	96.3%	86.1%	80.0%
本文所提方法	100.0%	99.5%	97.0%	87.5%	85.0%

基于加权朴素贝叶斯的水质数据分类研究

Research on water quality data classification based on weighted Naive Bayes

在线阅读

PDF下载

可视化

摘要/Abstract

引用本文

使用本文

图/表 12

参考文献 20

相关文章 0

Metrics

推荐阅读 0