T-SNE Based on Simulated Plant Growth Optimization Algorithm with t-Distribution Parameters

Abstract

SNE is a nonlinear dimension reduction method in the field of manifold learning in machine learning. In the field of high dimensionality, there will be some “noise” because the amount of de-space data is too large, that is, interference with the required data, so we need to reduce the dimension of these data. Although we obtain less information than high-dimensional, some features of the obtained data will become obvious for better observation and use. The loss function is KL divergence. The traditional method is solved by gradient descent method (GD) or stochastic gradient descent method (SGD) this method has some defects. In order to solve these defects, a t-SNE (t-distributed stochastic neighborhood embedding) algorithm is selected to simulate plant growth optimization algorithm.