Scatter/gather browsing communicates the topic structure of a very large text collection

Pirolli, Peter; Schank, Patricia; Hearst, Marti A.; Diehl, Christine

doi:10.1145/238386.238489

Cited by 149 publications

(120 citation statements)

References 9 publications

Supporting

Mentioning

119

Contrasting

Unclassified

Order By: Relevance

“…DataMountain [14] allows users to spatially layout web thumbnails in a pseudo-3D plane; their project extends the work by highlighting related documents. Related documents are computed with a fixed similarity metric (unlike our dynamic learning system), based on either a previous subject's categorizations, or a cosine distance between document word vectors.…”

Section: Spatial Layout For Document Collectionsmentioning

confidence: 99%

Helping Users Sort Faster with Adaptive Machine Learning Recommendations

Drucker

Fisher

Basu

2011

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Abstract. Sorting and clustering large numbers of documents can be an overwhelming task: manual solutions tend to be slow, while machine learning systems often present results that don't align well with users' intents. We created and evaluated a system for helping users sort large numbers of documents into clusters. iCluster has the capability to recommend new items for existing clusters and appropriate clusters for items. The recommendations are based on a learning model that adapts over time -as the user adds more items to a cluster, the system's model improves and the recommendations become more relevant. Thirty-two subjects used iCluster to sort hundreds of data items both with and without recommendations; we found that recommendations allow users to sort items more rapidly. A pool of 161 raters then assessed the quality of the resulting clusters, finding that clusters generated with recommendations were of statistically indistinguishable quality. Both the manual and assisted methods were substantially better than a fully automatic method.

show abstract

Section: Spatial Layout For Document Collectionsmentioning

confidence: 99%

Helping Users Sort Faster with Adaptive Machine Learning Recommendations

Drucker

Fisher

Basu

2011

Lecture Notes in Computer Science

View full text Add to dashboard Cite

show abstract

“…For example, there are more than 20 entries in Wikipedia for different well-known individuals under the name of Jim Gray 13 and 74 entries for Michael Smith 14 . Consequently, the diversity of search results has increased to the point that we must consider alternative presentations, providing additional structure to flat lists so as to effectively minimize browsing effort and alleviate information overload [45,91,125,20]. Over the years clustering has been accepted as the most promising alternative.…”

Section: A Framework For Personalized and Collaborative Clustering Ofmentioning

confidence: 99%

“…Many clustering algorithms have been proposed [45,91,125,126,127,59,117,65]. In the industry, well-known cluster-based commercial search engines include Clusty 15 , iBoogie 16 and CarrotSearch 17 .…”

Section: A Framework For Personalized and Collaborative Clustering Ofmentioning

confidence: 99%

“…Scatter/Gather [45,91] was an early cluster-based document browsing method that performs post-retrieval clustering on top-ranked documents returned from a traditional information retrieval system. The Grouper system [125,126] (retired in 2000) introduced the well-known Suffix Tree Clustering (STC) algorithm that groups Web search results into clusters labeled by phrases extracted from snippets.…”

Section: Clustering In Irmentioning

confidence: 99%

See 1 more Smart Citation

Rapid Exploitation and Analysis of Documents

Buttler¹,

Andrzejewski²,

Stevens³

et al. 2011

View full text Add to dashboard Cite

“…It attempts to make it easy to evolve a collection, e.g., expanding it by query-by-example operations or limiting it by applying a filter. Scatter/Gather (Pirolli et al, 1996) supports the browsing of large collections of text, allowing users to iteratively reveal topic structure and locate desirable documents. VIKI system (Marshall et al, 1994) lets user organize collections of items by arranging them in 2D space.…”

Section: Information Workpacesmentioning

confidence: 99%

Social Data Mining to Improve Bioinspired Intelligent Systems

Ochoa¹,

González²,

Castro³

et al. 2008

Data Mining in Medical and Biological Research

View full text Add to dashboard Cite

Social data mining systems enable people to share opinions and obtain a benefit from each other's experience. These systems do this by mining and redistributing information from computational records of social activity such as Usenet messages, system usage history, We report here on social data mining applications, systems that mine information from the structure and content of web pages and provide an exploratory information workspace interface. We carried out experiments that yielded positive answers to both evaluation questions. First, a number of automatically computable features about web sites do a good job of predicting expert quality judgments about sites. Second, compared to popular web search interfaces, the Topic Shop interface to this information lets users select significantly higher quality sites, in less time and with less effort, and to organize the sites they select into personally meaningful collections quickly and easily. We conclude by discussing how our results may be applied and considering how they touch on general issues concerning quality, expertise, and consensus The motivation for the social data mining approach goes back at least to Vannevar Bush's As We May Think essay. Bush envisioned scholars blazing trails through repositories of information and realized that these trails subsequently could be followed by others. Everyone could walk in the footsteps of the masters. In our work, we have formulated a similar intuition using the metaphor of a path through the woods. However, this metaphor highlights the role of collective effort, rather than the individual. A path results from the decisions of many individuals, united only by where they choose to walk, yet still reflects a rough notion of what the walkers find to be a good path. The path both reflects history of use and serves as a resource for future users. Social data mining approaches seek analogous situations in the computational world.Researchers look for situations where groups of people are producing computational records (such as documents, Usenet messages, or web sites and links) as part of their normal activity. Potentially useful information implicit in these records is identified, computational techniques to harvest and aggregate the information are invented, and visualization techniques to present the results are designed. Thus, computation discovers and makes explicit the "paths through the woods" created by particular user communities. And, unlike ratings-based collaborative filtering systems (Resnick et al., 1994), social data mining systems do not require users to engage in any new activity; rather, they seek to exploit user preference information implicit in records of existing activity. The "history-enriched digital objects" line of work (Hill et al., 1992) was a seminal effort in this approach. It began from the observation that objects in the real world accumulate wear over the history of their use, and that this wear -such as the path through the woods or the dog-eared pages in a paperback book or the smudges on...

show abstract

Scatter/gather browsing communicates the topic structure of a very large text collection

Cited by 149 publications

References 9 publications

Helping Users Sort Faster with Adaptive Machine Learning Recommendations

Helping Users Sort Faster with Adaptive Machine Learning Recommendations

Rapid Exploitation and Analysis of Documents

Social Data Mining to Improve Bioinspired Intelligent Systems

Contact Info

Product

Resources

About