A Semi-Supervised Learning Algorithm Based On Lie Group And Application

Semi-supervised learning has been taken much attention in recent years as a new machine learning method.In this paper,we introduce Lie group to semi-supervised Learning and put forward semi-supervised learning based on Lie group.The main research results are concluded as follows:1.Introduction of semi-supervised learning;2.The models of semi-supervised learning based on Lie group are given,including the algebra model and the geometric model;3.The algorithm of semi-supervised learning based on GL(n)(SSLA-GL(n))and its use in the prediction of activity of steroid drugs;4.The algorithm of semi-supervised learning based on parameter Lie group (SSLG-PLG)and its use in the dimensionality reduction and classification of UCI medical dataset;5.For drug toxicity pretiction,we give a simulation system based on SSLA-PLG to predict drug toxicity and validate it with tox dataset.Through the work of this paper,on one hand,the content of semi-supervised learning is enriched,on the other hand,we use semi-supervised learning based on Lie group to medical data in order to realize dimensionality reduction and classification,as well as the prediction of medical activity and toxicity,this can extend the novel applies of the above proposed new methods.