| 因素空间理论下基点分类算法研究 |
| |
| 引用本文： | 蒲凌杰,,曾繁慧,,汪培庄,.因素空间理论下基点分类算法研究[J].智能系统学报,2020,15(3):528-536. |
| |
| 作者姓名： | 蒲凌杰 曾繁慧 汪培庄 |
| |
| 作者单位： | 1. 辽宁工程技术大学 理学院，辽宁 阜新 123000;2. 辽宁工程技术大学 智能工程与数学研究院，辽宁 阜新 123000 |
| |
| 摘 要： | 目前，基于因素空间理论的背景基提取算法计算过程复杂，初始化必须依赖各因素极值，基点数量提取冗余等原因，未能在应用中取得很好效果。为此，结合内点判别法和知识可继承、可扩展的思想，提出一种计算简单、初始化独立、基点数量小的改进的背景基提取算法。然后，利用改进的背景基提取算法构造出一种全新的数据分类算法-基点分类算法，基点分类算法以提取每一类样本的背景基为预测模型，再通过新定义的λ-背景基，优化预测模型。数值实验表明:基点分类算法原理简单、构造难度小、分类模型泛化能力强，预测能力准确率高，同时严格的模型限定区域又能为识别新类别提供新方法。
|
| 关 键 词： | 因素空间 背景基 背景基提取 λ-背景基λ-背景基 基点分类算法 识别新类别 数据分类 背景分布 背景关系 |
Base point classification algorithm based on factor space theory |
| |
| PU Lingjie,,ZENG Fanhui,,WANG Peizhuang,.Base point classification algorithm based on factor space theory[J].CAAL Transactions on Intelligent Systems,2020,15(3):528-536. |
| |
| Authors: | PU Lingjie ZENG Fanhui WANG Peizhuang |
| |
| Affiliation: | 1. College of Science, Liaoning Technical University, Fuxin 123000, China;2. College of Intelligent Engineering and Mathematics, Liaoning Technical University, Fuxin 123000, China |
| |
| Abstract: | At present, the background-based extraction algorithm based on factor space theory has not achieved good results when used in applications. Reasons for its inefficiency are the calculation process is complicated, initialization depends on the extreme values of each factor, and redundancy of the number of base points extracted. Therefore, combining the inner point judgment method and a novel idea, an improved background-based extraction algorithm with simple calculation, independent initialization, and a few number of base points is proposed. Using the improved background-based extraction algorithm, a new data classification algorithm, i.e., base point classification algorithm, is constructed. The algorithm extracts the background base of each type of sample as the prediction model and optimizes the prediction model through the newly defined λ-background base. Finally, numerical experiments show that the base point classification algorithm is simple in principle, easy in construction, strong in generalizing the ability of classification model, and high in accuracy of prediction ability. Moreover, strict-utility model can provide new methods for identifying new classes. |
| |
| Keywords: | factor space background base background base extraction λ-background baseλ-background base base point classification algorithm identify new classes data classification background distribution background relationship |
|
| 点击此处可从《智能系统学报》浏览原始摘要信息 |
|
点击此处可从《智能系统学报》下载全文 |
|
