Research On Fuzzy Clustering Mining Technology And Its Application In Automatic Raising Pigs

At present, along with the information industry third wave coming, the internet ofthings which incorporates all kinds of information technology, breaks through thelimitation of the Internet and has the “comprehensive perception, reliable transmission,intelligent processing” features has penetrated into the husbandry areas. Throughout thecharacteristics and content of the modern husbandry, the informatization is the importantmeans to reflect the modern husbandry characteristics and realize the modern husbandrycontent. The integration or intersection of the precision husbandry decision requirementsand the intelligent calculation method include correlation, classification, clustering,evaluation and prediction. Therefore, in order to meet the urgent request of the farmers forhusbandry information technology, applying the kernel fuzzy clustering algorithm into theprocess of raising pigs is of great significance.The fatten pigs slaughter is an important part in the breeding process of the fatten pigs,and it is very easy to appear the lower slaughter rate condition without the scientificguidance. So, providing the slaughter recommendation services for farmers is ofsignificance. According to the requirements for slaughter, the slaughter recommendationservices classify the fatten pigs and recommend the suitable fatten pigs for slaughter,which spares the complex information collection and analysis work of the farmers. In orderto get a reasonable pig groups classification results, using the kernel fuzzy clusteringmining technology to analyze the data of the pig groups, the degree of uncertainty of piggroups was got and the intermediary of the pig category was expressed, which make theslaughter recommendation services more scientific.The main work of the paper is as follows:1、First, the background and the significance of the research were introduced, and thesignificance and the present research situation of the husbandry informazation system weremainly discussed, the theoretical foundation of FKCM was introduced in detail, and theprinciple and the steps of FKCM algorithm was elaborated.2、According to the large quantities of computation when using the effectivenessevaluation FKCM to calculate the optimal cluster number, the subtraction clusteringalgorithm was used to determine the cap of the initial clustering number, and the timecomplexity of the algorithm was reduced. As the original FKCM algorithm not consideringthe contribution to the clustering result of each sample is different, theinitial algorithm was improved based on the relaxation restriction feature weighting method.3、Some common effective indexes were introduced, and these were extended to thekernel space. Then a new clustering validity function based on the kernel space wasproposed, and the theory was proved, and the final experiment results show that the indexfunction have the very good capability to judge the validity of the clustering and thepractical value. In the clustering analysis on the pig group data, first the sample weight wasdone on the pig group data feature, and the clustering was done on the weighted data usingthe improved FKCM, finally the slaughter recommendation on the pig group was givencombining the expert investigation method. This algorithm not only can consider thecontribution degree of each sample on clustering, but can adjust the clustering results withthe subjective experience. The experiment results show that the FKCM based on thecombining the above methods make a more actual clustering effect in fatten pig slaughter.4、Finally, a slaughter recommendation model was designed using the improved FKCM,and a fatten pig slaughter recommendation system in the.net environment was realizedbased on the model, and it was applied into the fatten pig slaughter recommendationintelligence service.