Study On Complications Of Rheumatoid Arthritis Based On Novel K-means Clustering Algorithm Combining Particle Swarm Optimization And Bacterial Foraging Optimization

The electronic medical records(Electronic Medic al Records) have accumulated a large number of medical diagnostic information. How to provide references for doctors by mining the information hidden behind the massive medical records has became a hot issue in the current stud ies. Therefore, it is of great s ignific anc e and broad prospects for the data mining of electronic medical records to find valuable rules, to provide scientific bas is for c linic al experts in diagnosis and treatment as well as improving its level.This paper firstly proposes a novel c lustering algorithm to optimize the initial c luster centers of k-means clustering algorithm by combining bacterial foraging algorithm(BFO) which has high global searc h ability with partic le swarm algorithm(PSO) whic h is of high local search ability, bas ed on the bas is of the fact that initial values of traditional k-means c lustering algorithm is sens itive and eas y to fall into local optimum solution. As a result, the chemotaxis of bacteria is simp lified as the process of searching for the optimal solution of the partic les in the partic le s warm, and then the bacterias are used to complete the further operation of reproduction and elimination-dispersal. The determination of treating optimal solution of the hybrid algorithm as the initial clustering center provides a solution against the disadvantages of the k-means algorithm. The results on standard date set Iris,Wine, Glass of UCI display that the accurac y and stability of this proposed K-BFOPSO algorithm is higher than that of the popular c lustering algorithms, which can, at the same time, solve complex optimization problems more effectively.The open sourc e data mining tool, namely Weka, contains ric h function systems and is easy to operate. Through embedding the improved clustering algorithms in the Weka data, data can be analyzed intuitively. Thus, the Weka system’s development environment structure, the interface specification and the process of adding new algorithm and implementation steps into spec ific methods are analyzed, whic h realizes its secondary development to use the improved clustering algorithm embedded into Weka to mine data.Combined with the electronic medical records of patients with rheumatoid arthritis in The Second Affiliated Hospital of Shanxi Medic al Univers ity, the complication of rheumatoid arthritis is analyzed by means of the improved clustering algorithm in Weka. Through the c leaning and integration of the electronic medic al record, the data is imported into Weka and the improved clustering algorithm is introduced. At first, us ing the improved clustering algorithm clustering analyzes the sex and age of the RA patients’ medical records data to find out the high inc idenc e of RA related to all ages. Due to the mutual restraint and coupling exis ting between RA induced complications, clustering results suggest that the complic ations associated with RA are the diseases of cardiovascular system, lung dis ease, digestive system disease, endocrine and nutritional and metabolic diseases. And some complic ations appear simultaneous ly with concomitant. Through the analys is of RA records, the combination and age distribution related complic ations associated with RA are obtained. It should be paid muc h more attention to inspect and prevent the cardiovascular diseases, lung diseases of RA patients than before. In addition, the results indicate that there exis t apparent trend in gender and age among those RA patients.