Swets EPM Training Explorit Research Accelerator Focus Deep

Скачать презентацию Swets EPM Training Explorit Research Accelerator Focus Deep

65fbdff6eda25c6a3dfe66650e7a1e0c.ppt

Количество слайдов: 57

Agenda 1. Goals 2. Background 3. Hosted vs. Installed 4. Features and Benefits 5. Customization 6. Connectors 7. Authentication 8. Administration 9. SLA 10. Pilots 11. The Future 12. Additional Resources 13. Questions & Answers © 2009 Deep Web Technologies, Inc. Proprietary & Confidential

About Deep Web Technologies. . . • Founded by Abe Lederman, a cofounder of Verity, 2002 • Pioneered federated search technology • Over $3 M in R&D • Production applications since 1999 • Based in Santa Fe, New Mexico • 22 person company with strong executive team © 2009 Deep Web Technologies, Inc. Proprietary & Confidential

Customers Include. . . • Boeing • Defense Technical Information Center • DOE Office of Scientific & Technical Information • George Mason University • Intel Corporate Library © 2009 Deep Web Technologies, Inc. Proprietary & Confidential • Missouri Digital Heritage • National Agricultural Library • Science. gov Alliance • Scitopia. org • Stanford University • UCSF Medical School • World. Wide. Science Alliance

Sites Running Explorit • Scitopia. org • World. Wide. Science. org • Science. gov • Nutrition. gov • Biznar • Mednar • Science. Research. com © 2009 Deep Web Technologies, Inc. Proprietary & Confidential

What is “Federated Search”? “Federated Search is an application or service that allows users to submit a real-time search in parallel to multiple, distributed information sources and retrieve aggregated, ranked and de-duplicated results. ” © 2009 Deep Web Technologies, Inc. Proprietary & Confidential

In Other Words… One Search, Many Sources Begin Search Subscription Sources News E-Books © 2009 Deep Web Technologies, Inc. Proprietary & Confidential Public Web Sources Library Catalogs Internal Sources

Federated vs. Discovery Services Federated Search Discovery Service Real-time search of multiple collections Multiple collections are indexed to one database Search can take up to 30 seconds to complete Search completes within 1 second New results are available instantly New results are available only after re-indexing Searches full text where possible Mostly indexes just metadata Search any collection regardless of publisher Search only collections the service subscribes to © 2009 Deep Web Technologies, Inc. Proprietary & Confidential

What is the Best Solution for a Customer? • Hosted, Uses Explorit UI • Licensed, Uses Explorit UI • Hosted, Accessed via Web Services • Licensed, Accessed via Web Services © 2009 Deep Web Technologies, Inc. Proprietary & Confidential

Explorit Web Services API • Robust standards-based API • Easy to integrate into own Institution’s portal • Highly scalable • Locally deployed or provided as Saa. S • Supports all Explorit functionality through a number of Web Services: – Search Service (asynchronous) – Collection Service – User Service – Alerts Service – Statistics Service • Web-based Administrative Console © 2009 Deep Web Technologies, Inc. Proprietary & Confidential

Requirements for Licensed Solutions • Explorit runs on Linux – (Cent. OS 5. 4, Suse Enterprise 11, Redhat Linux 5) • Explorit runs on Windows Server • Explorit requires Tomcat Application Server Version 5. 5 or later • Explorit requires My. SQL or SQL Server © 2009 Deep Web Technologies, Inc. Proprietary & Confidential

Major Advantages of an Explorit Solution • Rich, easy-to-use, AJAX-based Web 2. 0 Interface • Incremental display of search results • Sophisticated/powerful connector building capability • Rich set of search capabilities • Retrieval of 50 -100 or more results per source • Relevance ranking (Admin tunable) © 2009 Deep Web Technologies, Inc. Proprietary & Confidential

Major Advantages of an Explorit Solution (cont. ) • Smart clustering and grouping of results • Alerts delivered via email or RSS • Search. Builder which enables custom search pages to be created • Selection of search results and export through Direct Connect • Administrative Console including rich set of graphics-based metrics © 2009 Deep Web Technologies, Inc. Proprietary & Confidential

Explorit Scalability • Searches more collections in parallel than any other federated search • Handles large number of concurrent searches • Able to federate other federated search engines © 2009 Deep Web Technologies, Inc. Proprietary & Confidential

Explorit Search Features • Basic Search • Advanced Search • Grouping of Collections into 2 -level Menu • Boolean Searching (AND, OR, NOT) • Fielded Searching • Author Normalization • Date Range Searches • Stemming / Wildcard truncation © 2009 Deep Web Technologies, Inc. Proprietary & Confidential

Explorit Result Page Features • Incremental Results • Merged/Ranked Results • Sorting Results • Refine Results • Limiting Results to a Single Source • Smart Clustering • Selection of Results • Email / Printing of Results • Export of Results (see Direct Connect) © 2009 Deep Web Technologies, Inc. Proprietary & Confidential

Relevance Ranking – How it Works • Is based on density of search terms in title/snippet • Search terms at beginning/end of title boost score • Higher/lower weights can be assigned to sources • Higher weights can be assigned to more recent results • Weighting of proximity of words © 2009 Deep Web Technologies, Inc. Proprietary & Confidential

Deduplication • Configurable through Admin Interface • Deduplicate on a field or combination of fields • Assign deduplication order priority to sources Example: URL or “Title and Date” © 2009 Deep Web Technologies, Inc. Proprietary & Confidential

Explorit Add-on Products Alerts • Alerts – Users register and maintain Alerts Home Page where they can: o Create/Update/Delete Alerts o View up to 6 most current Alert results – Delivered daily, weekly or monthly o Can be setup to generate a table of contents – Delivered via email, RSS or ATOM – Maintains a database of past results seen © 2009 Deep Web Technologies, Inc. Proprietary & Confidential

Explorit Add-on Products Search Builder/Direct Connect • Search Builder – Create custom search boxes and search pages easily – Enable/disable collections and search fields – Integrates with Course Management software • Direct Connect – Integrates with Selections capability – Export result metadata – Seamless integration with Ref. Works and End. Note © 2009 Deep Web Technologies, Inc. Proprietary & Confidential

Supported Browsers • Internet Explorer Version 7. 0 or higher • Firefox Version 2. 0 or higher • Google Chrome • Safari Supports screen resolution of 1024 x 768 © 2009 Deep Web Technologies, Inc. Proprietary & Confidential

Explorit Customer Customization • Incorporate customer header/footer, including logo • Specify fields to search on • Specify formatting of results • Select fields and order for Smart Clustering • Advanced Search page • Results list page • Results email • Alerts • Search Builder © 2009 Deep Web Technologies, Inc. Proprietary & Confidential

Swets Branded Explorit interface • Swetswise Searcher – UI look-and-feel – Admin interface look-and-feel • Graphics • Advanced Search page • Results list page • Results email • Alerts • Search Builder © 2009 Deep Web Technologies, Inc. Proprietary & Confidential

Connectors – Access Methods • Public sources requiring no authentication • IP-based authentication • User name / password • An organization’s proxy server Coming: Use of individual username and passwords © 2009 Deep Web Technologies, Inc. Proprietary & Confidential

Z 39. 50 Connectors • • Robust Z 39. 50 connector that searches 20+ fields Z 39. 50 connector tailored to work with major catalogs such as: – – – – © 2009 Deep Web Technologies, Inc. Proprietary & Confidential Sirsi. Dynix – Unicorn Sirsi. Dynix – Horizon Sirsi. Synix - Classic Dynix Ex Libris – Voyager Ex Libris – Aleph Innovative – Millennium VTLS – Virtua Polaris Connector Monitoring

Connector Monitoring • Proactively monitor connectors • Performed by dedicated software maintenance engineers • Monitor: source health, speed, responsiveness and errors • Generally errors are discovered by our team before users ever notice a problem © 2009 Deep Web Technologies, Inc. Proprietary & Confidential

Explorit Proxy Server The Proxy Server provides links to documents for sources that require authentication to retrieve search results and to retrieve full text documents from search results. • The Proxy Server supports forward proxy configuration for searching sources that require IP Authentication. • The Proxy Server supports reverse proxy configuration for searching sources that require any kind of HTTP authentication (Basic and/or form based). © 2009 Deep Web Technologies, Inc. Proprietary & Confidential

Proxy Configuration • Proxy url and configuration information (username/password, etc. ) may be configured through Strata admin interface (connector parameter configuration). • Proxy result urls search result lists are populated with result links that pass through the reverse Proxy to the protected content. © 2009 Deep Web Technologies, Inc. Proprietary & Confidential

Proxy Configuration • Proxy result urls search result lists are populated with result links that pass through the reverse Proxy to the protected content. • Each set of result urls generated for a a specific user/search are cryptographically hashed and stored for only a server-configurable amount of time so that access to protected content may not be abused. © 2009 Deep Web Technologies, Inc. Proprietary & Confidential

Open URL – Link Resolving • Works with major link resolvers – (SFX, Serials Solutions, Openly) • Passes DOI or combination of: – Journal title or ISSN – Article Volume & Page Number © 2009 Deep Web Technologies, Inc. Proprietary & Confidential

Explorit Administration Console • Enable/disable connectors • Connector monitoring & testing • Usage metrics • Click-through statistics • Tune relevance ranking • Configure de-duplication © 2009 Deep Web Technologies, Inc. Proprietary & Confidential

Explorit Metrics • Graphics-based or tabular • Single day (hourly breakdown) or entire month • Downloadable to spreadsheet • Reports include: – Number of queries run – Number of results retrieved per source – Average time to retrieve results from a source – Average rank of results retrieved per source – Timeouts/errors by source – Searches run (query strings) Coming: COUNTER compliance © 2009 Deep Web Technologies, Inc. Proprietary & Confidential

Guidelines for Building Pilots • Prospect must have seen demo of Explorit • Prospect must have budget and be actively evaluating federated search • Pilots must be limited to 6 sources • Prospect must provide credentials / access to source before start of pilot development • Pilot should run for no more than 60 days • A pilot agreement between Swets and the customer is a good way to show customer “buy-in” © 2009 Deep Web Technologies, Inc. Proprietary & Confidential

Our Roadmap Application • Web-based setup of customers • Interactive catalog of available connectors with self installation • Theme-based UI design / customization • Localization of UI into multiple languages • Save search queries / history • Enhanced metrics reports (COUNTER compliant) © 2009 Deep Web Technologies, Inc. Proprietary & Confidential

And Beyond. . . • Multi-lingual federated search • Integration with Social Networks • Source Selection Optimization • Personal libraries • Much more… © 2009 Deep Web Technologies, Inc. Proprietary & Confidential

Explorit Sales Literature • Deep by Design – Company Overview • Think Deep – Explorit Research Accelerator Overview • Reach Deep – Unlimited Collections • Aim Deep – Search Builder • Mednar – Deep Web Research for Medical Professionals • Alerts Marketing Piece [Needs to be updated] • Direct Connect Piece [Needs to be developed] © 2009 Deep Web Technologies, Inc. Proprietary & Confidential

Additional Resources • Federated Search Blog • “Quality not Quantity” White Paper • Federated Search Primer • “How to Maximize Your Strategic Investment in Federated Search” White Paper [Needs to be updated] • “From Requirements to Implementation: Federated Search at the Intel Library” • 100 Federated Search Requirements: Questions to Ask Vendors © 2009 Deep Web Technologies, Inc. Proprietary & Confidential

Presentations by Deep Web Technologies • Not All Federated Search Engines are Created Equal • Taking the Library Back from Google • Discovering Federated Search • Science Research: Journey to 10, 000 Sources (SLA) © 2009 Deep Web Technologies, Inc. Proprietary & Confidential