Managing Large Data Sets Using Support Vector Machines

Hundreds of Terabytes of CMS (Compact Muon Solenoid) data are being accumulated for storage day by day at the University of Nebraska-Lincoln, which is one of the eight US CMS Tier-2 sites. Managing this data includes retaining useful CMS data sets and clearing storage space for newly arriving data by deleting less useful data sets. This is an important task that is currently being done manually and it requires a large amount of time. The overall objective of this study was to develop a methodology to help identify the data sets to be deleted when there is a requirement for storage space. CMS data is stored using HDFS (Hadoop Distributed File System). HDFS logs give information regarding file access operations. Hadoop MapReduce was used to feed information in these logs to Support Vector Machines (SVMs), a machine learning algorithm applicable to classification and regression which is used in this Thesis to develop a classifier. Time elapsed in data set classification by this method is dependent on the size of the input HDFS log file since the algorithmic complexities of Hadoop MapReduce algorithms here are O(n). The SVM methodology produces a list of data sets for deletion along with their respective sizes. This methodology was also compared with a heuristic called Retention Cost which was calculated using size of the data set and the time since its last access to help decide how useful a data set is. Accuracies of both were compared by calculating the percentage of data sets predicted for deletion which were accessed at a later instance of time. Our methodology using SVMs proved to be more accurate than using the Retention Cost heuristic. This methodology could be used to solve similar problems involving other large data sets.

What Business Reference Librarians Need to Know about Intellectual Property: Presentations sponsored by the Patent and Trademark Depository Library Association (PTDLA) at the American Library Association Annual Conference, New Orleans, June 25, 2006

Presentations sponsored by the Patent and Trademark Depository Library Association (PTDLA) at the American Library Association Annual Conference, New Orleans, June 25, 2006 Speaker #1: Nan Myers Associate Professor; Government Documents, Patents and Trademarks Librarian Wichita State University, Wichita, KS Title: Intellectual Property Roundup: Copyright, Trademarks, Trade Secrets, and Patents Abstract: This presentation provides a capsule overview of the distinctive coverage of the four types of intellectual property – What they are, why they are important, how to get them, what they cost, how long they last. Emphasis will be on what questions patrons ask most, along with the answers! Includes coverage of the mission of Patent & Trademark Depository Libraries (PTDLs) and other sources of business information outside of libraries, such as Small Business Development Centers. Speaker #2: Jan Comfort Government Information Reference Librarian Clemson University, Clemson, SC Title: Patents as a Source of Competitive Intelligence Information Abstract: Large corporations often have R&D departments, or large numbers of staff whose jobs are to monitor the activities of their competitors. This presentation will review strategies that small business owners can employ to do their own competitive intelligence analysis. The focus will be on features of the patent database that is available free of charge on the USPTO website, as well as commercial databases available at many public and academic libraries across the country. Speaker #3: Virginia Baldwin Professor; Engineering Librarian University of Nebraska-Lincoln, Lincoln, NE Title: Mining Online Patent Data for Business Information Abstract: The United States Patent and Trademark Office (USPTO) website and websites of international databases contains information about granted patents and patent applications and the technologies they represent. Statistical information about patents, their technologies, geographical information, and patenting entities are compiled and available as reports on the USPTO website. Other valuable information from these websites can be obtained using data mining techniques. This presentation will provide the keys to opening these resources and obtaining valuable data. Speaker #4: Donna Hopkins Engineering Librarian Renssalaer Polytechnic Institute, Troy, NY Title: Searching the USPTO Trademark Database for Wordmarks and Logos Abstract: This presentation provides an overview of wordmark searching in www.uspto.gov, followed by a review of the techniques of searching for non-word US trademarks using codes from the Design Search Code Manual. These codes are used in an electronic search, either on the uspto website or on CASSIS DVDs. The search is sometimes supplemented by consulting the Official Gazette. A specific example of using a section of the codes for searching is included. Similar searches on the Madrid Express database of WIPO, using the Vienna Classification, will also be briefly described.