自动化学报2025，Vol.51Issue(10)：2256-2268,13.DOI:10.16383/j.aas.c250286
大回归模型的自适应学习
Adaptive Learning of Large Regression Models
摘要
Abstract
With the rapid development of information technology and the continuous improvement of computation-al power and data collection capability,modeling complex scenarios using large parameter models has become a sig-nificant development trend.However,the learning problems of such models under general feedback inputs still re-main lacking in the control system field.In view of this,we design an online adaptive learning algorithm with ex-panding dimension for large regression models under saturated observations.This algorithm can automatically up-date both algorithm dimension and computation results as new data increases,enabling dynamic adjustment of learning outcomes and real-time prediction of outputs without storing historical data.Specifically,we prove the con-vergence of the proposed algorithm under general non-persistent excitation data conditions,making it applicable to general feedback control systems.Moreover,the good convergence of the prediction"regrets"of the proposed al-gorithm is also proved without requiring any excitation data conditions.Finally,we conduct judicial sentencing pre-diction experiments based on real-world criminal judgment data for intentional injury cases,validating the compu-tational efficiency and prediction accuracy of the proposed algorithm.关键词
自适应学习/大回归模型/收敛性理论/一般数据条件/饱和观测/司法量刑Key words
Adaptive learning/large regression models/convergence theory/general data conditions/saturated ob-servations/judicial sentencing引用本文复制引用
戴瑞芬,王芳,郭雷..大回归模型的自适应学习[J].自动化学报,2025,51(10):2256-2268,13.基金项目
国家自然科学基金(T2293773,72371145,12288201),国家重点研发计划(2024YFC3307200),山东省泰山学者专项经费(tsqn202211004)资助Supported by National Natural Science Foundation of China(T2293773,72371145,12288201),National Key Research and De-velopment Program of China(2024YFC3307200),and Special Funds for Taishan Scholars Project of Shandong Province,China,(tsqn202211004) （T2293773,72371145,12288201）
