細(xì)胞型膜系統(tǒng)在聚類算法中的研究
發(fā)布時(shí)間：2018-08-03 08:18
【摘要】：膜計(jì)算是一種新型的計(jì)算方式，研究者從活細(xì)胞的結(jié)構(gòu)和功能中獲得靈感，從而抽象出這種計(jì)算模型。膜計(jì)算具有分布式、不確定性、最大并行計(jì)算等獨(dú)特特征，，使其在現(xiàn)有的計(jì)算科學(xué)研究中顯示出突出的優(yōu)勢(shì)，現(xiàn)已證明很多膜計(jì)算模型具有與圖靈機(jī)等價(jià)的計(jì)算能力，甚至具備超越圖靈機(jī)的可能，因此膜計(jì)算在解決一些并行問題、優(yōu)化問題方面，具有很重要的研究意義，并在生物學(xué)、醫(yī)學(xué)、計(jì)算機(jī)圖形學(xué)、語言學(xué)、經(jīng)濟(jì)學(xué)、社會(huì)學(xué)、計(jì)算機(jī)科學(xué)等多個(gè)領(lǐng)域有著廣泛的應(yīng)用。目前各個(gè)領(lǐng)域的專家學(xué)者對(duì)膜計(jì)算的專注度越來越高，使其成為全球?qū)W術(shù)界的研究熱點(diǎn)。 隨著信息產(chǎn)業(yè)界的發(fā)展和當(dāng)前社會(huì)的不斷進(jìn)步，數(shù)據(jù)挖掘領(lǐng)域受到了全社會(huì)各個(gè)領(lǐng)域的極大關(guān)注。而作為數(shù)據(jù)挖掘重要分支的聚類分析，是解決數(shù)據(jù)挖掘問題的主要手段。聚類分析就是從大量的樣本對(duì)象中，發(fā)現(xiàn)不同的對(duì)象分布情況，從而將相似的對(duì)象劃分成類的過程。目前聚類分析方法已經(jīng)在很多不同領(lǐng)域得到應(yīng)用，例如神經(jīng)網(wǎng)絡(luò)、圖像處理、現(xiàn)代生物學(xué)、統(tǒng)計(jì)學(xué)。 基于膜計(jì)算的非確定型和最大并行性等特點(diǎn)，本文將其引入聚類算法中，借助P系統(tǒng)模型保證聚類質(zhì)量的同時(shí)提高數(shù)據(jù)集規(guī)模和運(yùn)算速度，主要研究?jī)?nèi)容包括以下三個(gè)方面：一是根據(jù)K凝聚層次聚類算法的特點(diǎn)，結(jié)合膜計(jì)算的優(yōu)勢(shì)，提出了一種基于細(xì)胞型P系統(tǒng)的K AGNES算法，通過輸入具有n個(gè)對(duì)象的數(shù)據(jù)集合、對(duì)象集合矩陣及聚類個(gè)數(shù)k，結(jié)合P系統(tǒng)的特性，來對(duì)P系統(tǒng)的膜結(jié)構(gòu)、膜內(nèi)對(duì)象、膜內(nèi)規(guī)則及規(guī)則優(yōu)先關(guān)系進(jìn)行設(shè)計(jì)和構(gòu)建，最終得到n個(gè)對(duì)象的k個(gè)分組。二是為解決了傳統(tǒng)的基于密度的聚類算法進(jìn)行區(qū)域查詢的繁瑣復(fù)雜問題，構(gòu)建了一個(gè)細(xì)胞型P系統(tǒng)來實(shí)現(xiàn)DBSCAN算法。將膜計(jì)算應(yīng)用在基于密度的聚類算法的實(shí)現(xiàn)上，是膜計(jì)算應(yīng)用的一個(gè)創(chuàng)新之舉。活性膜P系統(tǒng)是一種特殊的P系統(tǒng)，細(xì)胞膜的分裂規(guī)則使其能夠?yàn)橛?jì)算提供指數(shù)個(gè)計(jì)算空間，具有解決聚類問題的獨(dú)特優(yōu)勢(shì)。本文結(jié)合DBSCAN算法的特點(diǎn)和活性膜P系統(tǒng)的優(yōu)勢(shì)，提出了一種基于活性膜P系統(tǒng)DBSCAN算法，作為第三個(gè)研究重點(diǎn)，使其能夠在更短的時(shí)間內(nèi)完成聚類過程，提高聚類算法的效率。 隨著互聯(lián)網(wǎng)的普及，電子商務(wù)系統(tǒng)對(duì)商品經(jīng)濟(jì)的發(fā)展和消費(fèi)者的日常經(jīng)濟(jì)生活，產(chǎn)生了翻天覆地的影響。但隨著電子商務(wù)的飛速發(fā)展，商品種類和數(shù)量的急劇增長(zhǎng)，電子商務(wù)結(jié)構(gòu)也變得越來越復(fù)雜，顧客很難在電子系統(tǒng)繁多的商品存儲(chǔ)中，精確地找到自己想要的商品，于是電子商務(wù)推薦系統(tǒng)便應(yīng)運(yùn)而生。根據(jù)目前電子商務(wù)網(wǎng)站的現(xiàn)狀及商品推薦所存在的問題，本文商品推薦系統(tǒng)看作一個(gè)無向加權(quán)圖并將其轉(zhuǎn)換成DBSCAN聚類問題，最后使用活性膜P系統(tǒng)來實(shí)現(xiàn)。這種全新的商品推薦方法在一定程度上提高顧客的購(gòu)買率，增強(qiáng)企業(yè)的競(jìng)爭(zhēng)力。商品推薦問題的成功解決，也將有助于膜計(jì)算在現(xiàn)實(shí)應(yīng)用方面進(jìn)行更深入的研究。
[Abstract]:Membrane calculation is a new method of computing. Researchers derive inspiration from the structure and function of living cells, thus abstracting this calculation model. Membrane calculation has the unique features of distributed, uncertain, and maximum parallel computing, which shows prominent advantages in the current research of computational science. Many membrane computing models have been proved. With the computing power equivalent to the Turing machine, and even the possibility of surpassing the Turing machine, membrane computing has a very important research significance in solving some parallel problems and optimizing problems. It has extensive applications in many fields, such as biology, medicine, computer graphics, linguistics, economics, social science, computer science and so on. Experts and scholars in various fields have paid more and more attention to membrane computing, making it a research hotspot in the global academic field.
With the development of the information industry and the continuous progress of the current society, the field of data mining has attracted great attention from all fields of society. As an important branch of data mining, clustering analysis is the main means to solve the problem of data mining. The process of dividing similar objects into classes has been applied in many different fields, such as neural networks, image processing, modern biology, statistics.
Based on the characteristics of uncertainty and maximum parallelism of membrane computing, this paper introduces it into clustering algorithm, with the aid of P system model to ensure the quality of clustering and improve the size and speed of data sets. The main research contents include the following three aspects: first, according to the characteristics of the clustering algorithm based on K, combined with the advantages of membrane computing, it is proposed. A K AGNES algorithm based on cellular P system is introduced. By input of data sets with n objects, object set matrix and cluster number k, combined with the characteristics of P system, this paper designs and constructs the membrane structure of the P system, the inside object, the rules of the membrane and the rule priority relations, and finally obtains the K grouping of the n objects. Two is to solve the problem. The traditional density based clustering algorithm is a complicated and complicated problem of regional query. A cell type P system is constructed to implement the DBSCAN algorithm. The application of membrane computing to the implementation of density based clustering algorithm is an innovative approach to the application of membrane computing. The active membrane P system is a special P system, the cell membrane splitting rule makes It can provide an exponential computing space for computing, and has a unique advantage to solve the clustering problem. In this paper, based on the characteristics of the DBSCAN algorithm and the advantages of the active membrane P system, a DBSCAN algorithm based on the active membrane P system is proposed. As the third research focus, it can complete the clustering process in a shorter time and improve the clustering algorithm. Efficiency.
With the popularity of the Internet, the electronic commerce system has a great impact on the development of commodity economy and the daily economic life of consumers. However, with the rapid development of electronic commerce and the rapid growth of commodity types and quantities, the structure of electronic commerce has become more and more complex, and it is difficult for customers to store the various kinds of electronic goods. According to the present situation of the e-commerce website and the problems existing in the recommendation of the commodity, this article is regarded as an undirected weighted graph and converted it into a DBSCAN clustering problem. Finally, it is realized by using the active membrane P system. The method of commodity recommendation improves the customer's purchasing rate to a certain extent and enhances the competitiveness of the enterprise. The successful solution of the problem of commodity recommendation will also help to make more in-depth research on the practical application of membrane computing.
【學(xué)位授予單位】：山東師范大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2014
【分類號(hào)】：TP38;TP311.13
[Abstract]:Membrane calculation is a new method of computing. Researchers derive inspiration from the structure and function of living cells, thus abstracting this calculation model. Membrane calculation has the unique features of distributed, uncertain, and maximum parallel computing, which shows prominent advantages in the current research of computational science. Many membrane computing models have been proved. With the computing power equivalent to the Turing machine, and even the possibility of surpassing the Turing machine, membrane computing has a very important research significance in solving some parallel problems and optimizing problems. It has extensive applications in many fields, such as biology, medicine, computer graphics, linguistics, economics, social science, computer science and so on. Experts and scholars in various fields have paid more and more attention to membrane computing, making it a research hotspot in the global academic field.
With the development of the information industry and the continuous progress of the current society, the field of data mining has attracted great attention from all fields of society. As an important branch of data mining, clustering analysis is the main means to solve the problem of data mining. The process of dividing similar objects into classes has been applied in many different fields, such as neural networks, image processing, modern biology, statistics.
Based on the characteristics of uncertainty and maximum parallelism of membrane computing, this paper introduces it into clustering algorithm, with the aid of P system model to ensure the quality of clustering and improve the size and speed of data sets. The main research contents include the following three aspects: first, according to the characteristics of the clustering algorithm based on K, combined with the advantages of membrane computing, it is proposed. A K AGNES algorithm based on cellular P system is introduced. By input of data sets with n objects, object set matrix and cluster number k, combined with the characteristics of P system, this paper designs and constructs the membrane structure of the P system, the inside object, the rules of the membrane and the rule priority relations, and finally obtains the K grouping of the n objects. Two is to solve the problem. The traditional density based clustering algorithm is a complicated and complicated problem of regional query. A cell type P system is constructed to implement the DBSCAN algorithm. The application of membrane computing to the implementation of density based clustering algorithm is an innovative approach to the application of membrane computing. The active membrane P system is a special P system, the cell membrane splitting rule makes It can provide an exponential computing space for computing, and has a unique advantage to solve the clustering problem. In this paper, based on the characteristics of the DBSCAN algorithm and the advantages of the active membrane P system, a DBSCAN algorithm based on the active membrane P system is proposed. As the third research focus, it can complete the clustering process in a shorter time and improve the clustering algorithm. Efficiency.
With the popularity of the Internet, the electronic commerce system has a great impact on the development of commodity economy and the daily economic life of consumers. However, with the rapid development of electronic commerce and the rapid growth of commodity types and quantities, the structure of electronic commerce has become more and more complex, and it is difficult for customers to store the various kinds of electronic goods. According to the present situation of the e-commerce website and the problems existing in the recommendation of the commodity, this article is regarded as an undirected weighted graph and converted it into a DBSCAN clustering problem. Finally, it is realized by using the active membrane P system. The method of commodity recommendation improves the customer's purchasing rate to a certain extent and enhances the competitiveness of the enterprise. The successful solution of the problem of commodity recommendation will also help to make more in-depth research on the practical application of membrane computing.
【學(xué)位授予單位】：山東師范大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2014
【分類號(hào)】：TP38;TP311.13
【參考文獻(xiàn)】
相關(guān)期刊論文 前8條
1 蔡穎琨,謝昆青,馬修軍;屏蔽了輸入?yún)?shù)敏感性的DBSCAN改進(jìn)算法[J];北京大學(xué)學(xué)報(bào)(自然科學(xué)版);2004年03期
2 龍真真;張策;劉飛裔;張正文;;一種改進(jìn)的Chameleon算法[J];計(jì)算機(jī)工程;2009年20期
3 戚正偉;尤晉元;;基于細(xì)胞膜演算的Web服務(wù)事務(wù)處理形式化描述與驗(yàn)證[J];計(jì)算機(jī)學(xué)報(bào);2006年07期
4 潘林強(qiáng);張興義;曾湘祥;汪雋;;脈沖神經(jīng)膜計(jì)算系統(tǒng)的研究進(jìn)展及展望(英文)[J];計(jì)算機(jī)學(xué)報(bào);2008年12期
5 張葛祥;潘林強(qiáng);;自然計(jì)算的新分支——膜計(jì)算[J];計(jì)算機(jī)學(xué)報(bào);2010年02期
6 郭曉娟;劉曉霞;李曉玲;;層次聚類算法的改進(jìn)及分析[J];計(jì)算機(jī)應(yīng)用與軟件;2008年06期
7 陳燕俐;洪龍;金達(dá)文;朱梧i
本文編號(hào)：2161157
本文鏈接：http://sikaile.net/kejilunwen/jisuanjikexuelunwen/2161157.html
最近更新
教材專著
