Creating Conceptual Access

PORT <port-peer-review>


Summary?    (6821)

Describes experiment to combine tools to create conceptual clusters of e-mail archives    (6825)

Especially Latent Semantic Analysis.
References: phmartin@gu.edu.au (Jun 19, 2002)

Clusters evaluated and refined manually    (6835)

References: phmartin@gu.edu.au (Jun 19, 2002)

Refined clusters may be used as input for another phase of automated clustering    (6841)

References: phmartin@gu.edu.au (Jun 19, 2002)

Background info?    (6931)

Uta Priss's work distinguishing associative and formal structures    (6936)

"associative structures"?    (6947)

Fuzzy    (6956)

Probabilistic    (6960)

Bottom-up    (6964)

Complex    (6968)

Depend on local context    (6972)

Examples?    (6997)

Latent Semantic Indexing    (7003)

"formal structures"?    (6951)

Precise    (6977)

Top-down    (6981)

Use symbolic logic    (6985)

Examples?    (7010)

WebKB    (7015)

Hypothesis?    (7021)

Knowledge representation should combine both, because human reasoning uses both    (7027)

cdent's approach to augmentation?    (7681)

Analyze large task    (7686)

Break it into smaller tasks    (7690)

Determine which tasks may benefit from automation    (7694)

Try some automation    (7698)

Keep automation where it works    (7702)

Attempt at combining associative and formal structures    (7038)

Similar to Uta Priss's work.
References: upriss@indiana.edu (Jun 19, 2002)

Faceted knowledge organization as their formal representation    (7045)

Important contribution to PORT    (7053)

General comments?    (6564)

Description too general    (6568)

References: phmartin@gu.edu.au (Jun 19, 2002)

Terms used without precision, definitions, or examples    (6574)

E.g. "cluster", "facet", "access structure", "coding messages".
References: phmartin@gu.edu.au (Jun 19, 2002)

Author agrees    (7089)

Excuse?    (7095)

Limited time    (7100)

Deadlines more flexible than thought    (7108)

Figures not helpful    (6582)

References: phmartin@gu.edu.au (Jun 19, 2002)

Do not show term/concept/facet    (6586)

No term/concept/facet determined at that stage    (7119)

Tool designed to help with this.

Nature of relations between nodes not explicit    (6591)

No references to use of Formal Concept Analysis    (6634)

How would tools/techniques/output described in this paper compared to that produced to FCA tools like Richard Cole's Email Concept Analysis tool?
References: phmartin@gu.edu.au (Jun 19, 2002)

Clustering techniques?    (7181)

Human evaluation process needs to be more explicit    (8063)

References: ademoor@kub.nl (Jun 26, 2002)

Stress augmentation metamodel    (8070)

Possibly with a figure.
References: ademoor@kub.nl (Jun 26, 2002)

Figures print badly    (8078)

Node numbers not readable.
References: ademoor@kub.nl (Jun 26, 2002)

Specific comments    (8118)

Potential utility of each method will emerge from side-by-side analysis    (8090)

"By enabling side-by-side comparison of various methods of semantic analysis, the means by which each method clusters and the potential utility of each method will emerge."

How to envision this?    (8099)

Criteria for comparing utility?    (8103)

"We built a testbed for evaluating tools."    (8125)

Isn't method about organizing knowledge, not evaluating tools?    (8131)

"extraction of recommended resources"    (8145)

What kind of resources and extraction?    (8151)

From where and to what purpose?    (8156)

Traversal function will help evaluators attribute facets to messages    (8163)

"We contemplate that this traversal function will allow
comparisons that will assist human evaluators attribute facets to the messages."

How is traversal function used/integrated in attribution process?    (8172)

(031)    (8191)

Do you have experience w/ coding practices?    (8195)

Types of coding other than for importance?    (8199)

Sources of coding terms?    (8203)

How resolve coding conflicts?    (8207)

Paper's hypothesis?    (6646)

Classic document indexing techniques can generate meaningful clusters of data    (6656)

References: phmartin@gu.edu.au (Jun 19, 2002)

Paper does not support this    (6664)

Does not produce precise answers to precise questions    (6671)

Only knowledge-base techniques do this. The downside is that knowledge bases have to be built manually.

Valid?    (7255)

Depends on your goals    (7270)

Goals?    (7276)

Find precise answer to precise question    (7261)

Learn generally, without precise goals    (7284)

Faceting better for this.

Classic indexing better for precise questions    (7300)

True statement of authors' hypothesis?    (7129)

Not according to cdent    (7134)

cdent: Project is effort to create tools to augment process of faceted classification    (7150)

Not replace, augment.
References: cdent@burningchrome.com (Jun 24, 2002)

Poorly worded in abstract    (7166)

What tools are used?    (6598)

Classic document indexing tool    (6609)

References: phmartin@gu.edu.au (Jun 19, 2002)

Does not produce anything "conceptual" to exploit    (6616)

Different/original from other indexing tools?    (6624)

Tool in progress    (7060)

Accept any input indicating clustering of documents    (7066)

Support evaluating and labelling clusters    (7077)

Relevance to PORT?    (7227)

PORT's purpose is to evaluate tools that may be useful    (7231)

References: cdent@burningchrome.com (Jun 24, 2002)

Authors' tool is for evaluating tools    (7248)

First step in multistage inquiry.
References: cdent@burningchrome.com (Jun 24, 2002)

Thought about extending work to use conceptual ontologies rather than terms?    (8239)

References: ademoor@kub.nl (Jun 26, 2002)

What happens after humans have manually labelled terms?    (8245)

References: ademoor@kub.nl (Jun 26, 2002)

Do discovered relations become available for automated processing in next analysis?    (8250)

References: ademoor@kub.nl (Jun 26, 2002)