SEMANTIC CLUSTERING AND SPATIAL HYPERTEXT CMSC-676 - Pranav Ganore - - PowerPoint PPT Presentation

SEMANTIC CLUSTERING AND SPATIAL HYPERTEXT

CMSC-676

• Pranav Ganore

INTRODUCTION: Spatial Hypertext

■ Hypertext is text which contains links to other texts. ■ The term was coined by Ted Nelson around 1965. ■ What is its most famous implementation ? ■ What is Spatial Hypertext ? And why it emerged ? – Widespread use of hypertext led to readers getting lost or confused. – NoteCard’s (Browser Card’s) – network map rather than document viewer. – gIBIS , Aquanet – visual networks containing typed links and nodes. – Move from document centered hypertext systems to Map based hypertext systems. – Relationship between nodes & Node P Proximity. – VI VIKI/VK VKB, , StoryS ySpace, , Ti Tinde derBox

Benefits of Spatial Hypertext.

It takes advantage of people's considerable visual recognition and intelligence Visual Memory & Pattern Recognition Enables the expression of relation variations It facilitates constructive ambiguity It supports emerging problem-solving strategies It reduces overhead in communicating with others

INTRODUCTION: Semantic Clustering

Sem emanti tic Clus usteri ering group

• ups sem

emanti tically eq equi uivalen ent t sea earch queri ueries es — wor

• rds

ds, phra rases es and d sen enten ences es — into

• clus

uster ers based ed on

• n mea

eaning. The e higher er th the e num umber er of

• f

ques uesti tion

• ns, wor
• rds

ds and d phra rases es with th a similar r mea eaning, th the e grea reater er th the e clus uster er. When en is Sem emanti tic Clus usteri ering us used? ed?

Unsatisfactory search results ? – customers react negatively !! No precise/incorrect answers for clusters - Serve as good indicators of missing information.

Generative Semantic Clustering in Spatial Hypertext

Need to discover methods for generating semantically meaningful document layouts to help people organize information according to their own significant criteria. A Generative System – Ideally collects articles & elements that are most relevant to the user and organize these elements to facilitate browsing, understanding & discovering. Information Triage

• The process of sorting through relevant

materials and organizing them to meet the needs of the task at hand.

Processes involved:

Inform

• rmation
• n Tri

riage defined as the process of sorting through relevant materials and

• rganizing them to meet the

needs of the task at hand. Individual & Context specific Approach Quantifying distinguishing useful features.

CombinFormation A mixed initiative system

https://ecologylab.net/combinFormation/about.html

System for browsing, searching, and collecting information elements in an information space. Initiatives : System’s generative actions & User’s direct manipulation actions. Proactively collects information from information sources. Concurrently generates a Spatial Hypertext.

Generative Semantic Clustering in Spatial Hypertext

Placement Algorithm Feature Extraction Semantic Clustering : Rel elatedn ednes ess Poten ential Interactive Mechanisms : Latching & Re-cluster Operations

Conclusions

■ Suitability of “Relatedness Potential”. ■ Organizes information in Spatial Hypertext with less efforts ■ Can be used to segregate or aggregate the information elements from different elements. ■ Can be used to translate an unorganized collection of bookmarks into semantically clustered spatial hypertext . ■ Preliminary evaluations showed participants preferred the clustered documents.

Thank You!!

References

■ Generative Semantic Clustering in Spatial Hypertext. Andruid Kerne, Eunyee Koh, Vikram Sundaram, J. Michael Mistrot Interface Ecology Lab Center for Study of Digital Libraries | Computer Science Department Texas A&M University, College Station, TX 77843, USA ■ Card, S.K., Mackinlay, J.D., Shneiderman, B. Readings in Information Visualization. Using Vision to Think. Morgan Kaufmann Publishers, Inc, San Francisco, CA, 1999 ■ Spatial Hypertext: An Alternative to Navigational and Semantic Links http://cs.brown.edu/memex/ACM_HypertextTestbed/papers/37.html ■ A Spatial Hypertext Wiki for Architectural Knowledge Management | Carlos Solís, ShyWiki Organization, Mexico City, Mexico | csolis@shywiki.org Nour Ali, Muhammad Ali Babar, Lero, University of Limerick, Limerick-Ireland {Nour.Ali, Muhammad.AliBabar}@lero.ie ■ Web Visualization of Temporal and Spatial Health Data from Smartphone App in Smart and Connected Community (SCC) | Sharmin Afroz, Electrical and Computer Engineering, University of Memphis, Memphis, TN 38452 USA, safroz@memphis.edu | Bashir I. Morshed, Electrical and Computer Engineering, University of Memphis, Memphis, TN 38152 USA, bmorshed@memphis.edu ■ https://natematias.com/mapViewer/note.php?id=3271618269 ■ http://www.eastgate.com/storyspace/index.html ■ https://www.csdl.tamu.edu/ensemble/?page_id=123/ ■ https://ecologylab.net/combinFormation/about.html ■ https://www.inbenta.com/en/technology/ai-nlp/semantic-clustering/ ■ https://en.wikipedia.org/wiki/Hypertext