Web Mining Seminar

CSE 450Web Mining SeminarSpring 2008MWF 11:10­12:00pm Maginnes 113Instructor: Dr. Brian D. DavisonDept. of Computer Science & EngineeringLehigh Universitydavison@cse.lehigh.eduhttp://www.cse.lehigh.edu/~brian/course/webmining/Course Objectives To gain a background in web mining techniques To become proficient at reading technical papers To gain knowledge of important current web mining research To gain experience presenting technical material To learn to write critical reviews of research papers To explore a research project in some depth and write and present a technical paper summarizing that workSpring 2008Web Mining Seminar2Teaching materials  Required Text: Web Data Mining: Exploring Hyperlinks, Contents and Usage data. By Bing Liu, Springer, ISBN 3-450-37881-2.  Optional Text: Data Mining: Practical Machine Learning Tools and Techniques, 2nd Ed. By Witten and Frank, Morgan Kaufmann Papers: Most (perhaps al ) available onlineAuthor's homepages Citeseer/ResearchIndex Google ScholarACM Digital LibraryIEEExploreSpring 2008Web Mining Seminar3Seminars are less formal We have a small class Introduce yourselves!Spring 2008Web Mining Seminar4Introduction toWeb MiningSpring 2008Web Mining Seminar5What is data mining? Data mining is also called knowledge discovery and data mining (KDD) Data mining is extraction of useful patterns from data sources, e.g., databases, texts, web, images, etc.  Patterns must be: valid, novel, potential y useful, understandableSpring 2008Web Mining Seminar6Classic data mining tasks Classification:mining patterns that can classify future (new) data into known classes.  Association rule miningmining any rule of the form X → Y, where X and Y are sets of data items. E.g., Cheese, Milk→ Bread [sup =5%, confid=80%] Clusteringidentifying a set of similarity groups in the data Sequential pattern mining:A sequential rule: A→ B, says that event A will be immediately followed by event B with a certain confidenceSpring 2008Web Mining Seminar7What is web mining? The process of discovering knowledge from web page content, hyperlink structure, and usage data Builds on existing data and text mining techniques, but adds many new tasks and algorithms Three types, based on sources of data (often combined in practice): Web structure mining Web content mining Web usage miningSpring 2008Web Mining Seminar8Importance of web data miningThe web is unique! Amount of information is huge and stil growing, on almost any topic, and changes continuously No single editorial control: significant variations in quality, much duplication, and data formats vary widely Significant information is linked (within and between web sites) Web reflects a virtual society --- interactions among people, organizations, and automated systems, no longer limited by geographyThe Web presents chal enges and opportunities for miningSpring 2008Web Mining Seminar9Importance of web data mining Online organizations generate a huge amount of data How to make best use of data? Knowledge discovered from web data can be used for competitive advantage. Online retailers (e.g., amazon.com) are largely driven by data mining. Web search engines are information retrieval (text mining) and data mining companies Web surfers/searchers need tools to find, recommend, organize, and extract useful information from the WebSpring 2008Web Mining Seminar10