机器学习算法比较

Algorithm Comparison of Machine Learning

如今,机械学习在数据挖掘、图像处理、自然语言处理以及生物特征识别等领域的应用已十分广泛。在机器学习中有一种"无免费午餐(NFL)"的定理,它指出没有任何一个算法可适用于每个问题,尤其是与监督学习相关的。因此,应尝试多种不同的算法来解决问题,同时还要使用"测试集"对不同算法进行评估,并选出最优者。笔者基于机器学习的发展,对几种常见算法优劣进行了研究分析,并讨论了其发展前景。

Nowadays,machine learning has been widely used in data mining,image processing,natural language processing and biometric recognition.There is a theorem of"no free lunch(NFL)"in machine learning,which points out that no algorithm can be applied to every problem,especially in relation to supervised learning.Therefore,we should try many different algorithms to solve the problem.At the same time,we should use"test set"to evaluate different algorithms and select the best one.Based on the development of machine learning,the advantages and disadvantages of several common algorithms are studied and analyzed,and their development prospects are discussed.