摘要
本文从数值型数据、分类型数据、文本与图像这四大最常见的数据源出发,尝试讲解部分常用特征工程方法,然后介绍了降维和特征选择这两大特征工程主题中的通用方法。本文从特征工程的视角看待机器学习问题,希望读者能更全面的了解机器学习的流程。
Starting from the four most common data sources of numerical data,categorized data,text and image,this paper attempts to explain some common feature engineering methods,and then introduces the general methods of dimension reduction and feature selection.This paper describes machine learning from the perspective of feature engineering,hoping that readers can have a more comprehensive understanding of the process of machine learning.
作者
顾亚文
GU Ya-wen(Jinken College of Technolog,Institute of Information Engineering and Digitial Art,Nanjing Jiangsu 211156)
出处
《数字技术与应用》
2020年第2期217-218,共2页
Digital Technology & Application
基金
江苏省2017年高校“青蓝工程”优秀青年骨干教师培养基金:苏财教[2017]183号。
关键词
常用特征工程
数据处理
降维
特征选择
commonly used feature engineering
data processing
dimension reduction
feature selection
