Assignments for LIS 4930 730

[tgsclass2]

Assignment 13

Properties of Social Media

Scale:  the number of users/number of people the information can reach

Immediacy:  how quickly information is disseminated

Heterogeneity:  users/people of diverse backgrounds  

Duplication:  information is repeated or copied

Semi-structuredness:  informal/low quality information connected using metadata

Model

Small-worldness:  six-degrees of separation

Scale Free Property:  degree distribution that asymptotically follows a power law

Network Evolution:  changed over time by adding or removing

The Future of Information Retrieval in Social Media

There are some challenges with trying to organize information in social media so it can be found again later.  The assignment of metadata is usually left up to the users, but not every user would use the same metadata to describe the information.  There is not much of a standard that is used to regulate metadata besides the users' discretion.  This means it is more difficult to retrieve information because not all of it has proper metadata.  Certain metadata is used because enough users use it that it becomes popular.  This helps mitigate the problem, but a solution is yet to be found.  

[tgsclass2]

Assignment 12

International Children's Digital Library

This library has over 4,000 books that come in 59 different languages.  The metadata is in the book's original language, in English and is translated to different languages based on available volunteer translators.  The interface is colorful and picture-oriented, great for children to work with.  When using the search engine, the results are relevant specific to the query.  Truncation yields more results that are not necessarily relevant.  It appears that Boolean searching is not supported.  

World's Digital Library

This library has over 10,000 items that appear in their original languages.  The metadata is translated to seven languages.  The interface is simple and modern, making it easy for everyone to use.  While using the search engine, any item that had any connection to the query, even if it was part of another word, was returned.  The rules for relevance were very lose.  I believe this could be due the items appearing in their original languages in the database.  Truncation is useful for retrieving more results, but they are as relevant as stated before.  Again, with the loose search results, Boolean searching is not very efficient as they return results with the Boolean operators themselves.  

[tgsclass2]

Assignment 11

Semantic Web

The World Wide Web Consortium (W3C) is promoting the usage of common data file types on the Internet.  It is a framework that would allow data to be shared and reused because of the set standards.  

Ontology

As a framework that is used to represent knowledge,  ontology is used to define the types, properties, and relationships of entities.  It is used to organize information.  

Semantic Web and Online Library Catalog

Libraries are starting to convert their catalogs to conform with this framework for organizing data.  It is a big change from the systems they have been using and will require a lot of time and effort to convert the catalog over to the new system.  

XML Schema

Used to define data, the XML schema is a programming language used to tag information within certain categories.  This is done through hierarchical trees, with one or more categories fitting under another category to define the outer category.  

RDF Schema

Another programming language used to tag information, the RDF schema is similar to the XML schema.  The exception is that the RDF schema does not used hierarchical trees to define the information.  

OIL

Working in conjunction with the RDF schema, OIL used simple description logic to define the classes, properties, and hierarchy of data.  It is used in classifying data.  

Web Services

Web Services use web applications to connect devices, laptops, tablets, smart phones, to each other over the Internet.  These applications have code that allow them to work together over the different platforms and exchange and manipulate data.  

Semantic Web Services

A combination of semantic web and web services, semantic web services is something that does not fully exist yet, but is being continually pursued.  Using the framework of the semantic web, the organization of ontology, and the connectivity of the web services, semantic web services would streamline and greatly improve the functionality of the creation, transfer, and manipulation of data.  

Semantic Web Services and Libraries

All of this is important to libraries because libraries classify, organize, and catalog information and the semantic web services will greatly improve this process.  All of this coming together will improve the interaction between people and information.  

http://semanticweb.com/the-future-of-libraries-linked-data-and-schema-org-extensions_b35315

[tgsclass2]

Project Part 2

Abstract

Three factors were used to evaluate the results of four search queries from three different databases.  The databases were to be evaluated to determine their level of performance in terms of results returned to the user.   The four search queries were necessary to test for one on the three factors.  

Introduction

There are many methods and databases used to search for information.  The results will vary depending on the methods and databases used.  This makes which methods and databases that are chosen very important because the results will vary depending on the ones used to search for information.  

Results

Three databases were searched:  Google, library catalog, and my database containing  the information in the previous post titled Project Part 1.  I evaluated the results of four search queries with the building block approach, relevance, and aboutness.  

My initial search query was "ships".  This returned many results in Google and the library catalog, but only one in my database.  The results were relevant and were satisfactory in terms of aboutness.  

For the building block approach I continued my search of the three databases with the search queries "ship*", "ships AND space", "ships OR space".  

Under the query "ship*", Google returned many results that were not very relevant, but were satisfactory in terms of aboutness.  The library catalog returned similar results.  My database did not return anything more than previously.  

Under the query "ships AND space", Google returned results that were very relevant and were satisfactory in terms of aboutness.  The library catalog returned very few results that were not very relevant, but were satisfactory in aboutness.  My data base returned only one result, but a different one than previously as the previous result did not include space.  It was relevant and was satisfactory in terms of aboutness.  

Under the query "ships OR space", Google returned different results that were relevant and were satisfactory in terms of aboutness.  The library catalog return very few results that were not very relevant and was not very satisfactory in terms of aboutness.  My database returned both of the aforementioned results, both of which were relevant and were satisfactory in terms of aboutness.

Summary

Overall my database performed the best under each of the four search queries with the building block approach, relevance, and aboutness.  Google came in second  and the library catalog was last.  Judging by the results of the tests for relevance and aboutness for each of the queries, the results from my database were always relevant and were satisfactory in terms of aboutness, the results from Google were not always very relevant, but were almost always satisfactory in terms of aboutness, and the results from the library catalog were not very relevant at all, but was relatively satisfactory in terms of aboutness.  Using these two factors in evaluating for the building block approach, my data base returned the best results, but very few, Google returned very good results, and the library catalog returned unsatisfactory results.  

________________________________________________________

                        |  Google  |  library catalog  |  my database

________________________________________________________

Relevance       |    5254    |         5321         |     5555

________________________________________________________

Aboutness       |    5544    |         5542        |      5555

________________________________________________________

Building Block  |       4       |            3           |         5

________________________________________________________

[tgsclass2]

Assignment 10

Multimedia

Includes:  images, audio, video, animation, interactivity, etc. 

These types of files are difficult to retrieve because of their innate nature primarily non-text based content.  Search engines typically use keywords to retrieve files, but that would not work on these types of media because the keyword search cannot process information that is not text.  This lead to two different types of approaches being created to be able to retrieve multimedia files.  

Description Based Retrieval vs. Content Based Retrieval of Still Images

The description based approach is reliant upon outside sources for descriptions of the file.  This includes the file's physical attributes and subject attributes.  This information is more commonly known as metadata.  This gives the file keywords for the search engine to find.  

The content based approach gets its description of the file from the contents of the file.  The color, shape, and texture are used to describe the file for the search engines to be able to find it.  

TREC

The Text REtrieval Conference developed an approach to retrieving files that were in multiple languages.  The program took the search query and cross-referenced a file with translations of the file in other languages to be more accurate at retrieving information.  This was needed because of the different characters and grammar used in the writing styles of the different languages.  

[tgsclass2]

Assignment 9

Library Catalog Searching

Information is categorized, but the results are not always very relevant to the search query.  The language used to categorize the information and needed for search queries is very formal.  This means searching can be very simple, yielding exact results, but not always relevant in the intended manner the user wanted.  Using natural language can yield relevant results, but because of the way the database is constructed, it would be more beneficial to the user to follow the formal language protocol.  

Web Searching - Google PageRank

Pages are ranked based on the number of and the quality of those pages that are linked to the page.  This is combined with the relevance of the information on the page.  The search results are then ranked by relevance to the search query.  Sometimes brings up non-relevant results.  The search language needed is much like natural language and is easy to use.  This can make searching difficult because of how fluid language can be, but it usually yields more relevant results.  Using formal language would still yield relevant results, but there may be a smaller number of them.  

[tgsclass2]

Assignment 8

Users of Information Retrieval

Who:  Information seeking or information searching, distinct individuals; these individuals are grouped by certain characteristics to better understand how the different groups interact with information retrieval systems; these characteristics include:  age, occupation, culture, personality, etc.

Their needs:  to find representations (documents, etc.) of information that are relevant to their search query as to satisfy their information need

Cognitive Model for User Needs

The cognitive model is user-oriented, and concentrates on how the users of information retrieval systems satisfy their information needs.  It takes into account how the user thinks and processes information, unlike other models that focus on the information retrieval system itself.  The major drawback to this model is that knowledge of cognitive processes is fairly limited.  This means that there is a limited understanding of users' cognitive activities while information seeking or information searching.  

[tgsclass2]

Assignment 7

Boolean Search

Uses the operations AND, OR, and NOT in search queries to facilitate information retrieval.  

Pros:

well understood

easy to build

many are already in operation

flexible operations

Cons:

training needed to fully utilize operations

operations do not include some relationships

no system to rate importance of terms in the search query

no control over amount of query results

search query can return null or overload of results

Vector Processing

Items in the database are given a value, binary or weighted, to indicate importance and relevance to certain terms in the search query.  

Pros:

just choose search terms

results are rated for importance

results are returned in order of relevance

includes relevance feedback for improved results

Cons:

assumed independence of search terms

unable to identify synonyms or phrases

ratings for importance/relevance are subject to vary

[tgsclass2]

Assignment 6

Searching "Information and Data Visualization" in Google using:  

Boolean Search

The first things that showed up were ads about Data Visualization.  The first results were the Wikipedia pages about Data and Information. The next results were about Information.  After trying several iterations of the search term, the site "data.gov" was the only differing result; the other results remained the same whether using the AND or OR operands.  

Truncation

With this operand, the results differed between the different truncated words, and also differed from the Boolean searches.  The same kind of ads were still there.  Some of the same results for websites were the same when just information was truncated, when just visualization was truncated, and when both information and visualization were truncated.  When just visualization was truncated, the results were much more different than previous searches, including results about "Edward Tufte".  

Fuzzy Search

This worked really well.  The ads and results stayed the same regardless of how badly I spelled the search term, until I removed the m from Information.  Once I had removed the m, the results no longer pertained to information, but instead to fashion while still including the results about data visualization, even though they were misspelled as well.  

[tgsclass2]

Project Part 1

50 Data Records

Books:  Title; Author; Publication Company; Year

Blacksmith Without Flaws; Luna; Prophecy Productions; Z20

Doctor Without Duty; Daenerys; Worldwares; Z49

Defenders Of The Banished; Favnir; Dragonetworks; BH2134

Companions Without Honor; Hudraer the Eternal One; Sunlight Limited; Z35

Spies Of Demons; Darilth Gentleheart; Shadoworks; Z37

Humans And Witches; Corinth the Swift; Clover Limited; 398

Hope With Strength; Mianth the Dark; Alpinebit; Z52

Fate Of The River; Otaehryn; Crow Enterprises; Z35

Scared Of The Night; Kementh Firebreath; Dragonetworks; BH1638

Never Trust Technology; Arun; Auraland; Z52

Kings And Protectors; Rugarth Longtail; Crow Enterprises; 14

Alien Of Everywhere; Peoni Whifflewort; Shadoworks; Z22

Soldiers With A Spaceship; Robin Pumpkinfall; Prophecy Productions; Z34

Boys Of Darkness; Knight Gilmyn; Crow Enterprises; 375

Guests And Foreigners; Maie Darklily; Clover Limited; Z15

Medics And Volunteers; Sphiel Fernsun; Icecaproductions; Z43

Culling Of Our Ship; Miles Bluewhistle; Worldwares; Z35

Appearance Of The Flight; Stardust Daisywhisper; Crow Enterprises; 109

Darkness Of Orbital Flight; Florian Lotuswax; Shadoworks; Z27

Cultured By A Rise Of Machines; Brock Figlake; Banshee Arts; Z38

Butterfly Without Courage; Dusk Sugargrove; Clover Limited; 125

Woman Of The Bow; Octavia Amberberry; Crow Enterprises; Z9

Assassins Of My Imagination; Balthazar; Dragonetworks; BH18

Fairies Of Magic; Empress Janet; Worldwares; Z37

Lions And Ogres; Daemn; Crow Enterprises; 65

Lords And Vultures; Genevieve Fernbreath; Dragonetworks; BH167

Bow Of The Night; Onyx Thunderbriar; Prophecy Productions; Z23

Crypt Of Dread; Willow Bitterlight; Prophecy Productions; Z25

Welcome To The Chasm; Abigail Elmtree; Sunlight Limited; Z42

Sounds In The South; Auriana; Clover Limited; Z1

Phantom Of The Mountain; Alaneo Sugarvine; Banshee Arts; Z14

Girl Of The River; Amnayel Tarshishim; Freakwood Publications; Z38

Knights Of Destruction; Kindel Daevi; Icecaproductions; Z46

Phantoms Of Yesterday; Timber Mooncliff; Clover Limited; 167

Heroes And Swindlers; Amaliel; Nightfly Industries; Z7

Defenders And Girls; Lomes; Crow Enterprises; 34

Victory Without Duty; Baelemabuhn Traibale; Clover Limited; 432

Murder With A Goal; Uugone Ises; Icecaproductions; Z47

Taste Of The Elements; Anastasia Persephone Oldfield; Banshee Arts; Z2

Trust My Husband; Groiruse; Shadoworks; Z28

Cyborg Of Eternity; Mondione; Nightfly Industries; Z25

Doctor Of History; Wayne Hunter; Sunlight Limited; Z34

Zombies Without Courage; Yamnoun Coppercloak; Nightfly Industries; Z24

Armies Without A Goal; Matthew the Sly; Freakwood Publications; Z15

Planet Of The End; Phantom Selwyn; Crow Enterprises; Z56

Bane Of Doom; Romelete Rivin; Heroshade Corporation; Z53

Faith Of Dreams; Phystral; Dragonetworks; BH2565

Honor Of Nightmares; Lyla Crawford; Heroshade Corporation; Z51

Signs In My Home; Dekreda Flintbane; Crow Enterprises; 264

Calling Magic; Nix Iden VuFable; Dragonetworks; BH1423

[tgsclass2]

Assignment 5

Boolean Search

AND:  used to narrow search by looking for both terms together

OR:  used to widen search by looking for both terms together and separately

NOT:  used to narrow search by excluding certain terms from search

"  ":  used to narrow search by looking for the exact phrase

(  ):  used with NOT and OR to narrow search; 

        used with AND and OR to widen search

Truncation vs. Boolean Search

* or ? or +:  used to widen search by taking the place of parts of words

It is different from Boolean search in that it is able to search using partial words where the Boolean search uses full words.  

Fuzzy Search vs. Boolean Search

Google had more search results than the Library of Congress Authorities.  Google was able to handle fuzzy searching with the same results while the Library of Congress Authorities could not and instead asked the user to check spelling and punctuation or to try a different type of search.  Both yielded accurate results.  Google sorted the results by relevance; the Library of Congress Authorities was sorted alphabetically.  

[tgsclass2]

Assignment 4

Natural Language in IR

The language used to speak or write in everyday life is considered natural language.  It is used as search queries to search for information in search engines and databases.  When it is used in search queries, there is some trouble with ambiguity in the way some things are phrased or how words can have differing meanings with different context.  To combat the ambiguity, lists were created to be used by the database to filter words that are too commonly used or have multiple meanings or spellings, or words that are used frequently within the database.  

http://www.upf.edu/hipertextnet/en/numero-5/pln.html

Controlled Vocabulary in IR

As an opposite of sorts to natural language, controlled vocabulary is very predetermined and very limited.  Because different databases focus on certain topic areas, the vocabulary used must be created in a way that it can still be usable even though most entries in the database will have the same keywords.  The vocabulary is constructed from literary warrant or user warrant; meaning it the keywords are chosen from preexisting entries in the database and past queries made by users.  Three types of the most common controlled vocabulary are thesauri, classification schemes, and subject headings.  

http://marciazeng.slis.kent.edu/Z3919/1need.htm

Thesauri in IR

Used mostly for post-coordination, thesauri is a type of controlled vocabulary that is flexible, specific, and able to handle complex search queries.  Standard notations are used in thesauri to weed out false coordination in search queries and consist of USE and UF (used for) identify which of the terms in the search query are preferred, SN (scope notes) identifies which meaning of the term, NT (narrower term) and BT (broader term) identify the hierarchical relationships between the terms, and RT (related terms) identify the associations between the terms.  Thesauri are most commonly organised either alphabetically or hierarchically.  

http://www.allegrotechindexing.com/article02.htm

Classification Schemes in IR

Used mostly for pre-coordination, classification schemes is a type of controlled vocabulary that is based on a system that is created to organize information.  Hierarchical relationships are used more so than the associations between the information.  The information is organised alphanumerically, numerically, or alphabetically.  The most common classification schemes used are the Dewey Decimal System and the Library of Congress Classification.  

http://www.ipl.org/div/farq/deweyFARQ.html

[tgsclass2]

Assignment 3

Dublin Core (DC) vs Resource Description Framework (RDF)

DC has 15 elements that are used in metadata to describe networked resources.  RDF does not have specific elements that are used to describe the networked resources.  Instead it has a model that can be used to define any element the user wishes.  The model is comprised of what the object is, an attribute of that object, and the value of the attribute.  

Metadata in Social Media

http://moz.com/blog/meta-data-templates-123

Metadata and Multimedia

As it can be difficult to describe a picture or sounds with words, the metadata used to describe multimedia tell what it contains.  It will include the basic of what type of media (text, image, sound) and then go on to describe the content of the media itself (color, shape, pitch, etc.).  

[tgsclass2]

Assignment 2

Recorded Information vs Social Media

Recorded information may be more commonly referred to as recorded history.  It is specific events of the past that have been documented.  Whereas social media documents the daily life of the individual and their interactions with others.  

Indexing vs Categorization 

Indexing uses terms to organize information.  These keywords or phrases are used to store and retrieve information.  Categorization, on the other hand, has a system that places information into different categories denoted by a sequence of a combination of numbers and letters.  The sequence of numbers and letters are used to store and retrieve information.  As in these two categorization systems:  

Dewey Decimal Classification

Library of Congress Classification

Citations

Citations are bibliographic information that reference other works that were used in the making of a document.  

Information Representation

http://www.csi.ucd.ie/staff/jcarthy/home/Information.html

[tgsclass2]

Assignment 1

Information vs. Data

Information is comprised of data.  Data is what is gathered by the senses.  It is then organized and arranged in a way that we can then understand more wholly.  It is then considered information.  

Information Representation

It is the way information is presented.  It can refer to the format the information takes, such as text, image, video, and audio.  It also has to do with the categorization of the data in the information presented.  

Information Retrieval

It uses information representation to find information once it has been categorized and stored.  Search Queries are used to locate the information being sought after.