=Paper=
{{Paper
|id=Vol-1466/invited03
|storemode=property
|title=None
|pdfUrl=https://ceur-ws.org/Vol-1466/invited03.pdf
|volume=Vol-1466
}}
==None==
<pdf width="1500px">https://ceur-ws.org/Vol-1466/invited03.pdf</pdf>
<pre>
  Clarifying Lattice Structure by Data Polishing

                                    Takeaki Uno

                Institute of Informatics (NII) of Japan, Tokyo, Japan

Abstract. Concept lattice is made from many kinds of data. We want to use large
scale data for the construction to capture wide and deep meanings but the result of the
construction usually yields a quite huge lattice that is impossible to handle. Several
techniques have been proposed to cope with this problem, but to best of our knowledge
no algorithm attains good granularity, coverage, size distribution, and independence of
concepts at the same time. We consider this difficulty comes from that the concepts
are not clear in the data, so a good approach is to clarify the concepts in the data by
some operations. In this direction, we propose “data polishing” that modify the data
according to feasible hypothesis so that the concepts becomes clear.

</pre>