Research On Water Quality Assessment And Time Series Prediction Based On Decision Tree And Related Weights

Lake is an important part of the global water resources,and plays an important role in maintaining the balance of the ecosystem.However,due to the influence of various human factors,the eutrophication of lakes is becoming more and more serious,which has become the most prominent problem in all kinds of water quality hazards.Taihu as China's second largest freshwater lake,is also the largest lake in China's eastern coastal region,since the last century since 80 s,the water quality of Taihu mutation,almost every year to 90 s after the outbreak of a large area of blue-green algae.Until 2007 after the state of governance in Taihu,the problem of eutrophication has been controlled.In order to improve the water quality evaluation method,this paper establishes a decision tree model,taking Taihu as the research object,based on the water quality monitoring data of 1992-2006 when serious outbreak of blue-green algae based on Meiliang Bay,a typical lake area in Taihu.To determine the important factors affecting the water quality of Taihu lake eutrophication;secondly,revise the evaluation system for the grade of eutrophication in Taihu,and related to the weight based method,the monthly water quality of Taihu on 2013-2017 was evaluated;finally,in the time series analysis of the eutrophication evaluation index of Taihu in 2007-2017,the influence of the independent variable on the model fitting accuracy is considered,and the comparison of the two model statistics is carried out.,and the grade prediction and prediction of water quality in Taihu in 2018 is given.Based on the analysis and study of the water quality system in Taihu,the following main conclusions are drawn:(1)In the decision tree model of Meiliang Bay in Taihu Lake established in this paper,16 indicators are selected as input variables.The model fitting results show that the permanganate index,total phosphorus,total nitrogen,pH,and water temperature affected Meiliang Bay.The eutrophication of water is an important factor,and the threshold of each indicator under different water quality conditions is given.(2)Based on the analysis of the decision tree model of Meiliang Bay in Taihu Lake,this paper revises the standard system of water quality and nutritional grade evaluation in Taihu Lake,and determines four evaluation indexes and seven evaluation grade evaluation methods,and uses the method of calculating related weights.Based on the evaluation of the water quality of Taihu Lake from January 2013 to December 2017.The results show that the water quality in the five years is between light eutrophication and eutrophication.The water quality in February 2015,April,May,June and July is light eutrophication,the other months are eutrophication,and the water quality is better in five years.This method is consistent with the actual water quality in Taihu Lake.This shows that the revised method is more detailed and the selection of evaluation indicators is more theoretically meaningful.The assessment of nutritional status is more consistent with the actual situation of water quality.(3)In this paper,the time series models of Chla,CODMn,TP and TN in Taihu Lake are constructed.The model fitting types are: Chla:ARIMA(1,0,0)(1,0,0);CODMn:ARIMA(0,1,1)(0,1,1);TP: ARIMA(0,0,1)(0,0,0);TN: ARIMA(1,0,0)(0,1,1).At the same time,the four indicators are used as the external factors of the time series model to increase the model fitting.The results show that the model fitting accuracy is higher than the model without the independent variables.Finally,based on the assessment of the water quality and nutrition level of Taihu Lake as revised in this paper,the predicted values of each month and each index in 2018 are compared.Finally,the predicted value of the nutritional level of Taihu Lake from January to December of 2018 is obtained.Departments and research scholars provide reference value.