Isomap Based on Cuckoo Search Algorithm with t-distribution Parameters

Isomap borrows the idea of geodesics which is from differential geometry, hoping data can keep geodesic distance on the manifold after transforming into low dimensional space mapping. Isomap takes advantage of Gradient descent or Stochastic gradient descent to solve the problem in the end. However, Isomap is prone to get trapped in local minimum, and the result obtained is always a local optimal solution. In order to solving such kind of problem, this article comes up with Cuckoo search to optimize the solutions about figuring out manifold distance.