【摘要】：線性非高斯無(wú)環(huán)模型(LiNGAM)具有在沒(méi)有任何先驗(yàn)知識(shí)的情況下能夠僅僅從觀察數(shù)據(jù)中完整地識(shí)別因果網(wǎng)絡(luò)的優(yōu)勢(shì),這使得它得到了越來(lái)越多研究者的關(guān)注.然而,現(xiàn)有求解LiNGAM模型的算法中一部分存在對(duì)初始值敏感,容易陷入局部最優(yōu)解的問(wèn)題,一部分存在對(duì)于外生變量識(shí)別率低的缺陷.為此,提出了一種基于最大最小獨(dú)立性的因果發(fā)現(xiàn)算法.通過(guò)引入自適應(yīng)的獨(dú)立性判定參數(shù),根據(jù)此參數(shù)來(lái)找出與其余所有變量回歸得到的殘差都獨(dú)立的變量,即為外生變量.該算法不僅避免了傳統(tǒng)算法對(duì)獨(dú)立性值差異敏感而導(dǎo)致識(shí)別率低的問(wèn)題,而且也避免了不同數(shù)據(jù)集對(duì)固定獨(dú)立性參數(shù)敏感而導(dǎo)致無(wú)法識(shí)別的缺陷.將該算法應(yīng)用于虛擬網(wǎng)絡(luò)和真實(shí)網(wǎng)絡(luò)中,實(shí)驗(yàn)結(jié)果都表明,各種維度下該算法都優(yōu)于現(xiàn)有的其他算法.
[Abstract]:The linear non-Gao Si acyclic model (LiNGAM) has the advantage of only recognizing causal networks completely from observational data without any prior knowledge, which has attracted more and more researchers' attention. However, some of the existing algorithms for solving the LiNGAM model are sensitive to the initial value and easily fall into the local optimal solution, while others have the defect of low recognition rate for exogenous variables. Therefore, a causal discovery algorithm based on maximum and minimum independence is proposed. By introducing an adaptive independence decision parameter, the variables which are independent of the residuals obtained from the regression of all the other variables are found according to this parameter, that is, the exogenous variables. This algorithm not only avoids the problem of low recognition rate caused by the sensitivity of traditional algorithms to the difference of independence value, but also avoids the defect that different data sets are sensitive to fixed independence parameters and can not be recognized. The experimental results show that the algorithm is superior to other existing algorithms in different dimensions.
【作者單位】： 廣東工業(yè)大學(xué);佛山科學(xué)技術(shù)學(xué)院;
【基金】：NSFC-廣東聯(lián)合基金(U1501254) 國(guó)家自然科學(xué)基金(61472089,61572143) 廣東省自然科學(xué)基金(2014A030306004,2014A030308008) 廣東省科技計(jì)劃項(xiàng)目(2013B051000076,2015B010108006,2015B010131015) 廣東特支計(jì)劃(2015TQ01X140) 廣州市珠江科技新星(201610010101)
【分類號(hào)】：TP18

【相似文獻(xiàn)】

相關(guān)期刊論文 前1條

1 劉佳昆;周永權(quán);;一種最大最小螢光素值人工螢火蟲算法[J];計(jì)算機(jī)應(yīng)用研究;2011年10期

相關(guān)碩士學(xué)位論文 前3條

1 何凝;基于最大最小距離法的音樂(lè)節(jié)拍跟蹤算法研究[D];天津大學(xué);2014年

2 馬文華;改進(jìn)最大最小蟻群算法的研究及應(yīng)用[D];安徽大學(xué);2015年

3 張曉丹;WSN中基于改進(jìn)粒子群優(yōu)化算法的分簇拓?fù)渌惴ㄑ芯縖D];鄭州大學(xué);2017年

，

本文編號(hào)：2306704