Continuous piecewise linear functions play an import role in approximation, regression and classification, and the problem of their explicit representation is still an open problem. In this paper, we propose a new form of representation,thus it can be seen clearly that they are actually a special kind of neural networks. Then we improve a typical piecewise linear algorithm called hinging hyperplanes to solve the training problem. Simulation results show that the novel algorithm has a much better approximation effect than the previous algorithm.
展开▼
