Spatio-temporal Patterns Mining And Influencing Factors Analysis Of Municipal Solid Waste In China Using Spatial Statistics And Machine Learning

China is the largest developing country in the world and is undergoing rapid and largescale urbanization.However,as the economy continues to grow,the environment is under increasing pressure.With the rapid growth of municipal solid waste(MSW)generation,traditional management is facing challenges.As a result,the waste problem has become a major obstacle to the sustainable development of economy,society and environment in China.In order to improve the efficiency of waste management and speed up the construction of urban ecological civilization,differentiated waste management solutions must be adopted in different regions.Therefore,it is of great practical significance to fully mine the spatiotemporal patterns and analyze influencing factors of MSW in China.However,MSW has the characteristics of extensive sources,various types and a large quantity.Its distribution pattern is variable in time and space,and is affected by many factors.It is difficult for existing methods to fully integrate its time,space and attribute data for spatiotemporal pattern mining and influencing factors analysis.Therefore,spatial statistics and machine learning are using to mine the spatio-temporal patterns and analyze influencing factors of MSW in China,aiming to provide auxiliary decision support for the adoption of differentiated MSW management schemes in different regions.Firstly,the space time cube of MSW is constructed to integrate the multi-dimensional space-time characteristics of MSW such as time,space and attributes,which based on the data related to MSW generation of 267 cities in China from 2002 to 2019.Then,the spatial statistical method is applied to the space time cube to fully mine the spatio-temporal distribution pattern and evolution pattern of MSW through local spatial autocorrelation,emerging spatio-temporal hotspot analysis and time series clustering.Finally,the random forest regression method is used to analyze the influencing factors of MSW generation.The research work has drawn the following results and conclusions:(1)From the perspective of changing trends,the amount of MSW generation in most cities in China shows an increasing trend over time.Only some cities in the Northeast China have fluctuated or even decreased slightly due to population outflow and relatively slow economic development.From the perspective of spatial distribution,the amount of MSW generation generally shows a high distribution pattern in the provincial capital cities and the southeastern coastal areas,but a relatively low distribution pattern in other cities.Especially in Beijing,Shanghai,Guangzhou and their surrounding cities which with large scale of urban population,city size and the rapid economic development,the amount of MSW generation is far above the average level.(2)The High/High clustering and Low/High clustering patterns of MSW are mainly distributed in North China,East China and South China,the Low/Low clustering patterns are mainly distributed in the central and western regions,and the High/Low clustering patterns are mainly distributed in the capital city of the southwest region.The spatio-temporal hot spot evolution pattern of MSW is dominated by consecutive hot spots,and does not show a spatiotemporal cold spot pattern.The hot spot trends are mainly distributed in North China,East China and South China.The high value of MSW is mainly distributed in Beijing,Shanghai,Chongqing,Shenzhen,Guangzhou,Dongguan and other provincial capital cities and cities in the southeast coastal areas,and the growth rate of MSW generation in these locations increases relatively rapidly over time.(3)The area of built district(ABD)and the urban population(UP)have a greater impact on the amount of MSW generation,while the industrial structure(IS),per capita expenditure of urban households(PCEUH)and per capita GDP(PCGDP)have a relatively weak impact on it.Since 2015,the influence of ABD has exceeded the influence of UP,and has become the most influential factor of MSW.The main influencing factors of MSW are different in different regions.In North China,South China and Southwest China,the influence of ABD is the highest,while in Northeast China,East China,Central China and Northwest China,the influence of UP is the highest.In summary,the space-time cube can effectively integrate the multi-dimensional spacetime characteristics of MSW such as time,space and attributes,and the application of spatial statistical methods to the space-time cube can fully explore the spatio-temporal distribution pattern and evolution pattern of MSW,while the random forest regression method has exerted its advantages in the measurement of the importance of influencing factors,and can effectively analyze the influencing factors of MSW generation.The relevant research conclusions can provide auxiliary decision-making reference for the reduction,recycling and harmless management of MSW.