Research On Spatial Clustering And Visualization Technology Of Agricultural Information

With the rapid development of information technology,its related technology has been widely used in the field of agriculture that makes the agricultural information technology become the research focus of the industry currently.With the development of the Geo-information system and data acquisition technology,produced a lot of spatial data which are numerous and multidimensional and also difficult to management and analysis,however,many valuable and useful information is hidden behind in those data.In order to get useful information from these data,it needs an effective data analysis means to achieve this goal.Aiming at the above-mentioned problems,this article studies and application of the data mining and visualization technology for agricultural data,my main work in this paper are:1.A fuzzy C means algorithm based on combined weighted distance is proposed.On the basis of introduce the fuzzy C means algorithm in detail,combined the key feature of agricultural information,this article focusing on the problem that different attributes has different influences and contribution,Then extended the Euclidean distance,proposed a generalized weighted Euclidean distance and improved the calculation method of weight.This article has introduced the concept of combination weight which was composed by subjective weights,and using the PCA(principal component analysis)to obtain the subjective weights,meanwhile introducing the weight coefficient β,and then use the additive synthesis method to calculate the last weight.Following building the Lagrange function for objective function to calculate the new updating formula and designed the process of algorithm.By comparing with Fuzzy c-means algorithm and Weighted fuzzy C-means algorithm,this method is been proved to have higher efficiency and the objective function has faster convergence speed.Finally,the fuzzy clustering algorithm in this paper is proved that the effect is more in line with the actual situation by comparative analysis in the practical application of agriculture.2.A cluster visualization model is proposed.During the clustering in order to make the data can be displayed in intuitive way,we mapped the data into parallel coordinates,and decomposed the process of theK-means algorithm and integrating into visualization in some process.Set the save mechanism for intermediate result which can be used visual way effectively to display.Through using the interactive technology of parallel coordinates of the data shows that the analysis results more clearly.Finally in the Matlab,the method is proved to be more effective to get the information.3.To test and verify the effectiveness of the presented method,applied to the three aspects that the classification of agricultural economics,the analysis and evaluation of agricultural natural disaster,the flood disaster prediction.The fuzzy clustering of this paper has a good classification result of the agricultural economic type,the parallel coordinate visualization method based on clustering has certain practical effect in the evaluation and prediction of agricultural natural disaster,which can more effectively extract useful information in agriculture,and it provides a way of though for the development of agricultural information.