摘要
机器学习作为一门源于人工智能和统计学的学科,是当前数据分析领域重点研究方向之一。首先通过追溯机器学习起源和介绍不同算法在求解策略上的启发性思路,讨论五类机器学习的发展及其主要算法在评价方法和优化方式上的实现,进一步总结归纳各算法适用领域和算法优劣,最后指出各算法克服自身缺陷的最新进展和未来实现多算法融合的研究方向。
Machine learning is a discipline derived from artificial intelligence and statistics,and it has been one of the key research directions in the field of data analysis.This paper introduces the inspiring ideas of different machine learning algorithms in the strategy through their origins,and the realization of five tribes of machine learning and the main algorithms including evaluation function and optimization method.Then applicable fields of each algorithm and both advantages and disadvantages of the algorithm are summarized.Finally this paper points out the latest progresses of each algorithm to overcome its own defects and the future research direction of multi-algorithm fusion.
作者
李旭然
丁晓红
LI Xu-ran;DING Xiao-hong(School of Mechanical Engineering,University of Shanghai for Science and Technology,Shanghai 200093,China)
出处
《软件导刊》
2019年第7期4-9,共6页
Software Guide
关键词
机器学习
学习算法
集成方法
增强理论
元学习
machine learning
learning algorithm
ensemble method
reinforcement learning
meta learning
