本文關鍵詞： Web日志 瀏覽偏愛路徑 MapReduce Hadoop　出處：《湖南工業(yè)大學》2015年碩士論文　論文類型：學位論文

【摘要】：隨著互聯(lián)網(wǎng)發(fā)展帶來的數(shù)據(jù)爆炸,使得Web服務器積累了大量的日志數(shù)據(jù),如何從海里的Web日志中挖掘有價值的信息成為了目前的研究熱點之一。通過對Web日志進行有效分析和挖掘,進而發(fā)現(xiàn)用戶瀏覽偏愛路徑,既可以為優(yōu)化網(wǎng)站的拓撲結構提供參考,而且又能為企業(yè)制定更完善的營銷策略提供依據(jù)。本文對基于Hadoop的用戶瀏覽路徑挖掘技術進行了相關研究,所做工作主要包括以下三個方面。1.提出并實現(xiàn)了一種基于可信興趣度的用戶瀏覽偏愛路徑挖掘算法。在充分考慮用戶瀏覽頁面時對頁面的感興趣程度的前提下,提出了頁面興趣度的概念;結合用戶瀏覽路徑選擇因素、頁面放置位置和其他頁面對該頁面的鏈接原因及網(wǎng)站拓撲結構圖修正加權衡量標準,提出了可信選擇度的概念。并將可信選擇度和頁面興趣度綜合度量,得到可信興趣度指標。提出并實現(xiàn)了基于可信興趣度的用戶瀏覽偏愛路徑挖掘算法(MUPCDI)。2.提出并實現(xiàn)了基于MapReduce的可信興趣度用戶瀏覽偏愛路徑挖掘算法,該算法運行于Hadoop分布式集群環(huán)境中,能對海量用戶瀏覽偏愛路徑進行分析與挖掘。3.針對目標數(shù)據(jù)集,應用基于可信興趣度的用戶瀏覽偏愛路徑挖掘算法(MUPCDI)對可信興趣度算法的閾值、準確性和有效性進行了對比分析;同時應用基于MapReduce的可信興趣度用戶瀏覽偏愛路徑挖掘算法對分布式平臺的高效性進行了對比分析。以上工作表明,本文提出的可信興趣度算法對于挖掘用戶瀏覽偏愛路徑更為準確和有效;同時針對挖掘大數(shù)據(jù)集的Web日志,分布式環(huán)境下本文提出的基于MapReduce的可信興趣度算法挖掘用戶瀏覽偏愛路徑的效率遠高于單機環(huán)境下。
[Abstract]:With the data explosion brought by the development of the Internet, the Web server has accumulated a lot of log data. How to mine valuable information from Web logs in the sea has become one of the research hotspots at present. By analyzing and mining Web logs effectively, we can find the preferred path for users to browse. It can not only provide a reference for optimizing the topological structure of the website, but also provide the basis for the enterprise to formulate a more perfect marketing strategy. This paper has carried on the related research to the user browsing path mining technology based on Hadoop. The main work includes the following three aspects. 1. A user preference path mining algorithm based on trusted interest is proposed and implemented. This paper puts forward the concept of page interest degree, combines with the factors of user browsing path selection, the reason of page placement and the link of other pages to the page, and modifies the weighted measurement standard of website topology chart. In this paper, the concept of trusted selection is put forward, and the trust selection and page interest are comprehensively measured. A user browsing preference path mining algorithm based on trusted interest is proposed and implemented. 2. A user browsing preference path mining algorithm based on MapReduce is proposed and implemented. The algorithm runs in the Hadoop distributed cluster environment and can analyze and mine the preference paths of massive users. The user browsing preference path mining algorithm based on trusted interest is used to analyze the threshold, accuracy and validity of trust interest. At the same time, using the trusted interest degree user browsing preference path mining algorithm based on MapReduce, the efficiency of distributed platform is compared and analyzed. The trusted interest algorithm proposed in this paper is more accurate and effective for mining user browsing preference paths, while mining the Web logs of big data set. The MapReduce based trusted interest algorithm proposed in this paper is much more efficient than the single computer environment in mining user browsing preference paths.
【學位授予單位】：湖南工業(yè)大學
【學位級別】：碩士
【學位授予年份】：2015
【分類號】：TP311.13

【參考文獻】

相關期刊論文 前9條

1 崔曉靖;陳興蜀;曾雪梅;;基于站點結構和瀏覽時間的路徑補全算法[J];計算機工程與設計;2014年03期

2 柳平;李春青;姬嬋娟;;基于HDFS的云存儲架構模型分析[J];電腦知識與技術;2013年36期

3 王思寶;李銀勝;;基于Web日志挖掘用戶的瀏覽興趣路徑[J];計算機應用與軟件;2012年01期

4 程苗;;基于云計算的用戶瀏覽偏愛路徑挖掘算法[J];計算機工程與應用;2011年29期

5 吳晶;張品;羅辛;盛浩;熊璋;;門戶個性化興趣獲取與遷移模式發(fā)現(xiàn)[J];計算機研究與發(fā)展;2007年08期

6 郭巖,白碩,楊志峰,張凱;網(wǎng)絡日志規(guī)模分析和用戶興趣挖掘[J];計算機學報;2005年09期

7 邢東山,沈鈞毅,宋擒豹;從Web日志中挖掘用戶瀏覽偏愛路徑[J];計算機學報;2003年11期

8 邢東山,沈鈞毅,宋擒豹;用戶瀏覽偏愛模式挖掘算法的研究[J];西安交通大學學報;2002年04期

9 韓家煒,孟小峰,王靜,李盛恩;Web挖掘研究[J];計算機研究與發(fā)展;2001年04期

，

本文編號：1551799