Exploiting parallelism in a structural scientific discovery system to improve scalability

Galal, Gehad; Cook, Diane J.; Holder, Lawrence B.

doi:10.1002/(sici)1097-4571(1999)50:1<65::aid-asi8>3.0.co;2-g

Cited by 7 publications

(4 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To make algorithms scalable also requires using more computational power -i.e., faster CPUs. In some cases this may be enhanced by multiple CPUs; however, the algorithms then need to be redesigned to take advantage of the explicit parallelism in hardware (Galal, Cook, and Holder, 1999). Consideration of which algorithm to be used might also depend on the specific variables that are employed.…”

Section: Discussionmentioning

confidence: 99%

Benchmarking Data Mining Algorithms

Rajagopalan

Krovi

2002

Journal of Database Management

View full text Add to dashboard Cite

Data mining is the process of sifting through the mass of organizational (internal and external) data to identify patterns critical for decision support. Successful implementation of the data mining effort requires a careful assessment of the various tools and algorithms available. The basic premise of this study is that machine-learning algorithms, which are assumption free, should outperform their traditional counterparts when mining business databases. The objective of this study is to test this proposition by investigating the performance of the algorithms for several scenarios. The scenarios are based on simulations designed to reflect the extent to which typical statistical assumptions are violated in the business domain. The results of the computational experiments support the proposition that machine learning algorithms generally outperform their statistical counterparts under certain conditions. These can be used as prescriptive guidelines for the applicability of data mining techniques.

show abstract

Section: Discussionmentioning

confidence: 99%

Benchmarking Data Mining Algorithms

Rajagopalan

Krovi

2002

Journal of Database Management

View full text Add to dashboard Cite

show abstract

“…Before describing CRA in detail, we review existing automated text analysis approaches in the next section. Galal et al (1999) General content analysis approaches Humphrey (1999) Keyword frequency approaches Rice & Danowski (1993) WORDij (Danowski, 1992) TACT (Siemens, 1993) APPROACHES TO AUTOMATED TEXT ANALYSIS A time-tested method of dealing with data floods is computer processing, which will be the approach advocated here. Specifically, we believe that a computerized analysis can serve as a substitute (though not necessarily a replacement) for a complete reading of a text.…”

Section: "Be Careful What You Wish For"mentioning

confidence: 99%

Studying Complex Discursive Systems: Centering Resonance Analysis of Communication

Corman¹

2002

Human Communication Research

118

206

View full text Add to dashboard Cite

Scholars increasingly theorize about the power of communication to organize and structure social collectives. However, two factors threaten to impede research on these theories: limitations in the scope and range of existing methods for studying complex systems of communication and the large volume of communication produced by even small collectives. Centering resonance analysis (CRA) is a new text analysis method that has broad scope and range and can be applied to large quantities of written text and transcribed conversation. It identifies discursively important words and represents these as a network, then uses structural properties of the network to index word importance. CRA networks can be directly visualized and can be scored for resonance with other networks to support a number of spatial analysis methods. Following a critique of existing methodologies, this paper describes the theoretical basis and operational details of CRA, describes its advantages relative to other techniques, demonstrates its face validity and representational validity, and demonstrates its utility in modeling organizational knowledge. The conclusion argues for its applicability in several organizational research contexts before describing its potential for use in a broader range of applications, including media content analysis, conversation analysis, computer simulations, and models of communication systems.1986; Schiffrin, 1994) makes conversation and interaction analysis less common than perhaps it should be in the communication discipline. Moving to the small group level, the detailed study of interaction is more difficult and rare still, and at the multigroup (i.e. organizational or cultural) level it is virtually unheard of. As our emerging theories and models of communication grow in scope to embrace complex collective phenomena, we risk making them unworkable as guides for empirical research. Put simply, we worry that the existing body of communication research methods is incapable of handling the complexity being theorized in the discipline.In this article, we propose a general analytical framework called centering resonance analysis (CRA), a flexible means of representing the content of large sets of messages, and assisting in their analysis. We present CRA as a text analysis method, suited to studying formalized communication like reports, letters, memos, emails, and other written texts. However, as we illustrate below, CRA is also applicable to transcribed conversations. Developments in voice recognition technology promise the ability in the near future to "textualize" live conversation, making it amenable to CRA methods as well. In establishing the need for CRA, we ground our arguments in the concerns of organizational communication because this is where the need to study large volumes of communication is especially acute. However, this does not mean CRA is only useful in research on formal organizations. As we explain in the conclusion, CRA has many applications in the broader discipline to problems in rhetoric, mass c...

show abstract

“…Some important programs are Inexact graph matcher, Minimum Description Length, distributed MPI Subdue [4] and Subgraph Isomorphism.…”

Section: Subduementioning

confidence: 99%

Subdue

Ketkar

Holder

Cook

2005

Proceedings of the 1st International Workshop on Open Source Data Mining: Frequent Pattern Mining Implementations

Self Cite

View full text Add to dashboard Cite

A majority of the existing algorithms which mine graph datasets target complete, frequent sub-graph discovery. We describe the graph-based data mining system Subdue which focuses on the discovery of sub-graphs which are not only frequent but also compress the graph dataset, using a heuristic algorithm. The rationale behind the use of a compressionbased methodology for frequent pattern discovery is to produce a fewer number of highly interesting patterns than to generate a large number of patterns from which interesting patterns need to be identified. We perform an experimental comparison of Subdue with the graph mining systems gSpan and FSG on the Chemical Toxicity and the Chemical Compounds datasets that are provided with gSpan. We present results on the performance on the Subdue system on the Mutagenesis and the KDD 2003 Citation Graph dataset. An analysis of the results indicates that Subdue can efficiently discover best-compressing frequent patterns which are fewer in number but can be of higher interest.

show abstract

Exploiting parallelism in a structural scientific discovery system to improve scalability

Cited by 7 publications

References 15 publications

Benchmarking Data Mining Algorithms

Benchmarking Data Mining Algorithms

Studying Complex Discursive Systems: Centering Resonance Analysis of Communication

Subdue

Contact Info

Product

Resources

About