The Improved Transductive Support Vector Machine and Its Application to Environment Monitoring in Industrial Seaculture

The study on Transductive Support Vector Machine(TSVM) has made little progress since Vapnik put forth the concept in the late 1990s, as algorithm for TSVM optimization model can not be easily found. Here we try to transform the problem of TSVM optimization into an unconstrained one before constructing the smooth unconstrained optimization that has a kernel, and on the basis of which to devise a TSVM whose optimization problem is easier to solve to breakthrough the bottleneck in order to deepen the research into TSVM and apply TSVM to environment monitoring of industrial sea culture.