Application Of Support Vector Machines In Medcal Data

With the "big medical data","precise medical","individualized medical” and other medical concepts put forward, how to apply internet technology successfully to medical data, and to dig out valuable information quickly and efficiently is an urgent problem to be solved. At the same time, for patients, there is a need for efficient disease diagnosis.In order to solve the above problems, this paper applies the support vector machine classification technique to the sleep field, classifies the sleep quality of the patients, and puts forward the incremental learning model suitable for medical data according to the characterisics of medical data.This paper first introduces the background and characteristics of large medical data. On this basis, an incremental learning algorithm(FK-SVM) combilning fuzzy C-means(FCM) algorithm and generalized KKT(Karush-Kuhn-Tucker) condition is proposed. The algorithm filters the data in two directions: the historical sample point and the new sample point, The FCM algorithm is used to filter the historical sample points, the screening of the new sample points is based on the generalized KKT condition, which reduces the number of training samples, greatly improving the algorithm's incremental learning ability,under the promise of does not influence the correct rate, running time has been reduced.Secondly, based on the theoretical analysis of FK-SVM algorithm,the construction of the experimental platform is completed, selected two standard databases(Letter and Magic) on UCI as experimental data, using the standard SVM algorithm and ISVM algorithm as the contrast algorithm. The experimental results show that the accuracy of the standard SVM algorithm and FK-SVM algorithm is significantly higher than that of the IS VM algorithm in the case of full data training, while the ISVM algorithm and FK-SVM algorithm are about 1/4 of the standard SVM operation time. Therefore, FK-SVM algorithm has better performance than standard SVM algorithm and classical ISVM incremental learning algorithm, and has strong incremental learning ability.Finally, the FK-SVM incremental learning algorithm applied to the internet-assisted psychotherapy (IAPS) system, as well as complete the design and implementation of the system. The system is divided into four modules: registration log, report analysis, self-detection and decompression. FK-SVM algorithm in the report analysis module to train the sleep data and to achieve incremental learning, to complete the support vector machine algorithm in the application of medical data research.