| With the development of science and technology and social progress,the scale of all kinds of data is growing rapidly.Through further analysis and processing,massive data can release more internal potential value and promote social and economic development,which has become a research hotspot.At present,the evaluation system of forestry soil fertility data relies more on manual analysis,which makes it vulnerable to subjective factors.There are some problems,such as large filling error of fertility data,incompleteness of fertility evaluation dimension and so on.At the same time,the soil fertility of forest land is affected by many factors such as soil bulk density and organic matter content because of its complex nutrient composition,which makes the data characteristics of soil fertility changeable and difficult to evaluate.In view of the above problems,this paper puts forward a data processing method of forest land soil fertility,and designs and develops a forest land soil fertility evaluation system.The specific work is as follows:(1)Aiming at the problem of missing soil fertility data in forest land,a missing data imputation algorithm based on clustering is proposed.Firstly,the clustering algorithm is used to select the data with the greatest similarity in the same cluster to fill in,so as to reduce the error between the data and ensure the accuracy of the filled data value;At the same time,considering the difference between numerical data and character data,different filling calculation methods are given;Finally,outlier detection is used to detect and iteratively correct the filled data.In order to verify the feasibility of the method,the missing data imputation experiment is carried out on the soil fertility data of forest land.The experimental results show that the filling error of the proposed data imputation method is small,the accuracy is high,and the integrity of fertility data can be guaranteed.(2)Aiming at the problems of low accuracy and poor evaluation effect of the existing forest land soil fertility evaluation model,a forest land soil fertility evaluation model based on improved capsule network is proposed.Firstly,the dense convolution network is used to extract the features of soil fertility data,and the spatial attention mechanism is used to weight the passed features,so that the model can obtain more local features,and then the fertility features are transformed into vectors through the capsule network,so as to integrate the advantages of multiple network structures and optimize the fertility feature extraction process.Finally,the compression function in the capsule network is optimized to ensure the integrity of multiple features in the network model,which has a significant optimization effect.(3)The analysis and evaluation system of forest land soil fertility is designed and implemented.The system adopts the Spring Boot design evaluation mode and builds the system platform to provide intelligent services for forest land users.On the web side,it realizes many functions,such as user login,data collection and upload,filling of forest land soil fertility data,fertility evaluation and fertility analysis,so as to improve the human-computer interaction ability and sense of experience,and fully display the research content of the subject.Finally,a comparative experiment was carried out on the forest land soil fertility data set,and the evaluation of forest land soil fertility was completed according to the data analysis of13 key fertility attributes such as soil texture,organic matter and total nitrogen.The results show that the accuracy of the proposed fertility evaluation model can reach 98.7%,and the optimization effect is more obvious compared with the comparison model,which verifies the effectiveness and optimization of the proposed forest soil fertility evaluation model in fertility evaluation.At the same time,the proposed evaluation model is applied to the forest land soil fertility evaluation system,which has a certain practical application. |
PDF Full Text Request