计算机科学
特征选择
分类器(UML)
人工智能
支持向量机
决策树
阿达布思
模式识别(心理学)
遗传算法
机器学习
选择(遗传算法)
特征(语言学)
数据挖掘
统计分类
哲学
语言学
作者
Utpal Singh,Minakhi Rout
标识
DOI:10.1109/inc457730.2023.10263100
摘要
A crucial data pre-processing step in data mining is feature selection. Choosing a subset of potential features by removing attributes with essentially no predictive value and heavily related redundant features appears to be the main goal of feature selection. We proposed a model that uses GA (Genetic Algorithm) for choosing or selecting these potential features. our model takes the set of features from the input data and then based its processing it encode it with either 0 or 1 which indicates the exclusion or inclusion the that feature. We use the Breast Cancer dataset for comparing the accuracy of different classification algorithms before applying the Genetic Algorithm and after applying it. We apply this feature selection method to different classifiers like RandomForest, Decision Tree, AdaBoost, Linear SVM, and a few others. With the help of the selected features, the accuracy of the model increases by 2.09% which outperformed the accuracy of a classification model which uses RandomForest classifier. This paper describes how Genetic Algorithms (GAs) help in selecting potential features for different classifiers.
科研通智能强力驱动
Strongly Powered by AbleSci AI