An accelerator for the logistic regression algorithm based on sampling on-demand  

作　　者：Jiye LIANG Yunsheng SONG Deyu LI Zhiqiang WANG Chuangyin DANG 

机构地区：[1]Key Laboratory of Computational Intelligence and Chinese Information Processing of Ministryof Education,School of Computer and Information Technology,Shanxi University,Taiyuan 030006,China [2]College of Information Science and Engineering,Shandong Agricultural University,Tai'an 271018,China [3]Department of Systems Engineering and Engineering Management,City University of Hong Kong,Hong Kong 999077,China

出　　处：《Science China(Information Sciences)》2020年第6期222-224,共3页中国科学（信息科学）（英文版）

基　　金：supported by National Natural Science Foundation of China(Grant Nos.61876103,61432011);Project of Key Research and Development Plan of Shanxi Province(Grant No.201603D111014);1331 Engineering Project of Shanxi Province。

摘　　要：Dear editor,Training a logistic regression classifier is equivalent to solving a convex optimization problem,which is frequently solved with gradient descent(GD).As the gradient is computed using all the data,using GD is very time-consuming when the dataset is large[1,2].To accelerate the GD algorithm,stochastic gradient descent(SGD)estimates the gradient using only one training data point.

关 键 词：SOD logistic regression LOGISTIC 

分 类 号：O224[理学—运筹学与控制论] TP181[理学—数学]