多示例多標(biāo)記主動(dòng)學(xué)習(xí)方法的研究及應(yīng)用
發(fā)布時(shí)間：2019-05-09 13:21
【摘要】：主動(dòng)學(xué)習(xí)作為一種能夠解決傳統(tǒng)分類問題中樣本的標(biāo)記信息缺失問題的機(jī)器學(xué)習(xí)方法受到了科研及實(shí)際應(yīng)用領(lǐng)域的關(guān)注。多示例多標(biāo)記學(xué)習(xí)作為一種新型的機(jī)器學(xué)習(xí)方法,能夠處理現(xiàn)實(shí)任務(wù)中復(fù)雜的學(xué)習(xí)任務(wù),目前尚未看到主動(dòng)學(xué)習(xí)應(yīng)用于多示例多標(biāo)記學(xué)習(xí)任務(wù)中的研究。本文主要開展多示例多標(biāo)記主動(dòng)學(xué)習(xí)方法及其應(yīng)用的研究,主要的工作內(nèi)容如下:(1)設(shè)計(jì)了基于標(biāo)記排序的多示例多標(biāo)記主動(dòng)學(xué)習(xí)方法框架MIMLAL。對(duì)傳統(tǒng)主動(dòng)學(xué)習(xí)模型、選擇規(guī)則和多示例多標(biāo)記框架等相關(guān)內(nèi)容進(jìn)行研究,根據(jù)目前較為先進(jìn)的基于標(biāo)記的機(jī)器學(xué)習(xí)思想提出了多示例多標(biāo)記主動(dòng)學(xué)習(xí)方法框架MIMLAL,解決主動(dòng)學(xué)習(xí)任務(wù)計(jì)算量復(fù)雜信息容易丟失的問題。(2)提出兩種具體的多示例多標(biāo)記主動(dòng)學(xué)習(xí)算法MIMLAL-A和MIMLAL-R。MIMLAL-A為基于標(biāo)記空間最大均值差異的多示例多標(biāo)記主動(dòng)學(xué)習(xí)方法。傳統(tǒng)的基于標(biāo)記的分類方法使用關(guān)鍵示例的方法在主動(dòng)學(xué)習(xí)中不夠準(zhǔn)確。本文依據(jù)最大均值差異MMD的思想將單示例主動(dòng)學(xué)習(xí)價(jià)值函數(shù)LCI擴(kuò)展到多示例下得到多示例MILCI并依此提出多示例多標(biāo)記主動(dòng)學(xué)習(xí)方法MIMLAL-A。MIMLAL-R為基于示例預(yù)測(cè)值排序的多示例多標(biāo)記主動(dòng)學(xué)習(xí)方法。MILCI使用平均的思想會(huì)引入不相關(guān)示例的無效信息,對(duì)此本文提出對(duì)示例按預(yù)測(cè)值排序加權(quán)的RLCI價(jià)值函數(shù)并針對(duì)單一選擇規(guī)則對(duì)樣本區(qū)分度不夠的問題,基于結(jié)合RLCI和標(biāo)記選擇的新型選擇規(guī)則提出MIMLAL-R方法。(3)實(shí)現(xiàn)了MIMLAL-A和MIMLAL-R方法在蛋白質(zhì)功能預(yù)測(cè)問題上的應(yīng)用。本文首先將新物種蛋白質(zhì)功能預(yù)測(cè)任務(wù)抽象為多示例多標(biāo)記主動(dòng)學(xué)習(xí)問題,然后應(yīng)用多示例多標(biāo)記主動(dòng)學(xué)習(xí)方法進(jìn)行建模。在多個(gè)新物種的蛋白質(zhì)功能預(yù)測(cè)任務(wù)中,通過對(duì)比發(fā)現(xiàn),我們的多示例多標(biāo)記主動(dòng)學(xué)習(xí)方法取得了非常好的預(yù)測(cè)性能,其中MIMLAL-R在大多情況下性能最優(yōu)。
[Abstract]:As a kind of machine learning method which can solve the problem of missing marker information of samples in traditional classification problems, active learning has attracted much attention in the field of scientific research and practical application. As a new machine learning method, multi-example multi-tag learning can deal with complex learning tasks in real-world tasks. At present, active learning has not been applied to the study of multi-example multi-tag learning tasks. The main contents of this paper are as follows: (1) A multi-sample multi-tag active learning method framework MIMLAL. based on tag sorting is designed in this paper, and the main work is as follows: (1) A multi-example multi-tag active learning method framework based on tag sorting is designed. In this paper, the traditional active learning model, selection rules and multi-example multi-tag framework are studied. According to the advanced markup-based machine learning idea, a multi-example multi-tag active learning method framework, MIMLAL, is proposed. To solve the problem that the complex information of active learning task computation is easy to lose. (2) two specific multi-example multi-tag active learning algorithms MIMLAL-A and MIMLAL-R.MIMLAL-A are proposed based on the maximum mean difference in tag space. Multi-example multi-tag active learning method. The traditional tag-based classification method using key examples is not accurate enough in active learning. In this paper, we extend the single example active learning value function LCI to the multi-sample MILCI according to the idea of the maximum mean difference MMD. Accordingly, we propose the multi-sample multi-marker active learning method MIMLAL-A.MIMLAL-R as the example-based active learning method. Multi-example and multi-tag active learning method for value ranking. MILCI uses the idea of average to introduce invalid information about unrelated examples. In this paper, the RLCI value function weighted according to the predicted value of the example is put forward, and the problem of insufficient classification of the sample area by a single selection rule is put forward. Based on the new selection rules combined with RLCI and marker selection, a MIMLAL-R method is proposed. (3) the application of MIMLAL-A and MIMLAL-R methods in protein function prediction is realized. In this paper, the task of protein function prediction of new species is abstracted as a multi-example multi-tag active learning problem, and then a multi-example multi-tag active learning method is used to model the problem. In the prediction task of protein function in many new species, we find that our multi-example multi-marker active learning method has achieved very good prediction performance, among which MIMLAL-R has the best performance in most cases.
【學(xué)位授予單位】：南京郵電大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2017
【分類號(hào)】：TP181
本文編號(hào)：2472801
[Abstract]:As a kind of machine learning method which can solve the problem of missing marker information of samples in traditional classification problems, active learning has attracted much attention in the field of scientific research and practical application. As a new machine learning method, multi-example multi-tag learning can deal with complex learning tasks in real-world tasks. At present, active learning has not been applied to the study of multi-example multi-tag learning tasks. The main contents of this paper are as follows: (1) A multi-sample multi-tag active learning method framework MIMLAL. based on tag sorting is designed in this paper, and the main work is as follows: (1) A multi-example multi-tag active learning method framework based on tag sorting is designed. In this paper, the traditional active learning model, selection rules and multi-example multi-tag framework are studied. According to the advanced markup-based machine learning idea, a multi-example multi-tag active learning method framework, MIMLAL, is proposed. To solve the problem that the complex information of active learning task computation is easy to lose. (2) two specific multi-example multi-tag active learning algorithms MIMLAL-A and MIMLAL-R.MIMLAL-A are proposed based on the maximum mean difference in tag space. Multi-example multi-tag active learning method. The traditional tag-based classification method using key examples is not accurate enough in active learning. In this paper, we extend the single example active learning value function LCI to the multi-sample MILCI according to the idea of the maximum mean difference MMD. Accordingly, we propose the multi-sample multi-marker active learning method MIMLAL-A.MIMLAL-R as the example-based active learning method. Multi-example and multi-tag active learning method for value ranking. MILCI uses the idea of average to introduce invalid information about unrelated examples. In this paper, the RLCI value function weighted according to the predicted value of the example is put forward, and the problem of insufficient classification of the sample area by a single selection rule is put forward. Based on the new selection rules combined with RLCI and marker selection, a MIMLAL-R method is proposed. (3) the application of MIMLAL-A and MIMLAL-R methods in protein function prediction is realized. In this paper, the task of protein function prediction of new species is abstracted as a multi-example multi-tag active learning problem, and then a multi-example multi-tag active learning method is used to model the problem. In the prediction task of protein function in many new species, we find that our multi-example multi-marker active learning method has achieved very good prediction performance, among which MIMLAL-R has the best performance in most cases.
【學(xué)位授予單位】：南京郵電大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2017
【分類號(hào)】：TP181
【參考文獻(xiàn)】
相關(guān)期刊論文 前5條
1 吳建盛;;基于新型機(jī)器學(xué)習(xí)方法的蛋白質(zhì)功能預(yù)測(cè)與分析[J];信息通信;2012年05期
2 孫吉貴;劉杰;趙連宇;;聚類算法研究[J];軟件學(xué)報(bào);2008年01期
3 龍軍;殷建平;祝恩;趙文濤;;主動(dòng)學(xué)習(xí)研究綜述[J];計(jì)算機(jī)研究與發(fā)展;2008年S1期
4 趙悅;穆志純;;基于委員會(huì)投票選擇方法的主動(dòng)學(xué)習(xí)的研究[J];太原理工大學(xué)學(xué)報(bào);2006年04期
5 殷志祥;蛋白質(zhì)結(jié)構(gòu)預(yù)測(cè)方法的研究進(jìn)展[J];計(jì)算機(jī)工程與應(yīng)用;2004年20期
相關(guān)碩士學(xué)位論文 前1條
1 唐俊;基于多示例多標(biāo)記學(xué)習(xí)的手機(jī)游戲道具推薦[D];南京大學(xué);2015年
，本文編號(hào)：2472801
本文鏈接：http://www.sikaile.net/kejilunwen/zidonghuakongzhilunwen/2472801.html
最近更新
教材專著
