知識(shí)推送系統(tǒng)中一種基于多分類徑向基神經(jīng)網(wǎng)絡(luò)的知識(shí)匹配方法(英文)
發(fā)布時(shí)間:2022-07-11 19:59
聚焦知識(shí)匹配領(lǐng)域,開展提高產(chǎn)品設(shè)計(jì)中知識(shí)推送系統(tǒng)性能的探索性研究。傳統(tǒng)匹配算法需重復(fù)計(jì)算,導(dǎo)致響應(yīng)時(shí)間長,準(zhǔn)確性也有待提高。本文目標(biāo)是實(shí)現(xiàn)對(duì)設(shè)計(jì)者知識(shí)需求的快速響應(yīng),并提供優(yōu)質(zhì)知識(shí)推送服務(wù)。在改進(jìn)之前工作基礎(chǔ)上,研究實(shí)際操作中增強(qiáng)有限訓(xùn)練集的兩種方法:案例特征向量中振蕩特征權(quán)值和修正案例特征。此外,提出一種多分類徑向基神經(jīng)網(wǎng)絡(luò),可從知識(shí)庫中一次性匹配知識(shí)并保證推送結(jié)果準(zhǔn)確性。使用數(shù)控機(jī)床中導(dǎo)軌設(shè)計(jì)的訓(xùn)練集訓(xùn)練和測(cè)試該方法,實(shí)驗(yàn)結(jié)果表明增強(qiáng)訓(xùn)練集有效,本文提出的方法優(yōu)于其他匹配方法。
【文章頁數(shù)】:15 頁
【文章目錄】:
1 Introduction
2 Related works
2.1 Knowledge push
2.2 Knowledge matching
2.3 Artificial neural networks
3 Technical approach
3.1 Augmented training set
3.1.1 Oscillation of the feature weight s
3.1.2 Revision of the case feature key
3.2 Multi-classification radial basis function neural network
4 Experiments
4.1 Data
4.2 Methods
4.3 Evaluation metrics
4.4 Results and discussion
5 Conclusions
Contributors
Compliance with ethics guidelines
本文編號(hào):3658785
【文章頁數(shù)】:15 頁
【文章目錄】:
1 Introduction
2 Related works
2.1 Knowledge push
2.2 Knowledge matching
2.3 Artificial neural networks
3 Technical approach
3.1 Augmented training set
3.1.1 Oscillation of the feature weight s
3.1.2 Revision of the case feature key
3.2 Multi-classification radial basis function neural network
4 Experiments
4.1 Data
4.2 Methods
4.3 Evaluation metrics
4.4 Results and discussion
5 Conclusions
Contributors
Compliance with ethics guidelines
本文編號(hào):3658785
本文鏈接:http://sikaile.net/guanlilunwen/gongchengguanli/3658785.html
最近更新
教材專著