Air Quality Data Filling Based On Matrix Filling Method

In recent years,the problem of urban environmental pollution has become more and more serious with the continuous improvement of the level of urbanization,the rapid development of urban industry,and the rapid increase of urban population.People pay more and more attention to air quality issues,and higher requirements for air quality monitoring and data processing.In view of the lack of completeness of the obtained air quality data,we use the vertical matrix filling method to fill the air quality data which based on sparse representation.In order to compare the point data filling and curve data filling methods,take the PMM method as the common multiple interpolation method is used as the comparison method to obtain useful conclusions in different missing cases of data.First,we summarize the previous research results in this field and propose not to convert the data into curve data for filling.We introduce the knowledge background and related theory of matrix filling.Then we introduce the basic idea of sparse representation and the vertical matrix filling algorithm.Then,in order to compare the accuracy of multiple interpolation,SLI and SLR,we extract a complete air quality data as a sample for random deletion of data processing and filling.We compare the sample accuracy and clustering effect.After that,filling the application for actual air quality data,we determined that the data of air quality has function character and performed a low rank test at first,and then filled the actual data.We clustered the actual air quality data after filling.In the end,we combine and summarize the work of this paper,and point out the shortcomings and the next research direction.Based on the determined that data of air quality having the function character and low rank character,we considered that the methods of SLI and SLR have universality at first.Then we used these two methods to fill data of air quality that has uncertainty of missing rate.We got a good result.Basically,we achieved the object of our goal.