講演名 2004/11/27
Analysis of Hepatitis Dataset by Using Cl-GBI(Medical Active Mining)(Joint Workshop of Vietnamese Society of AI, SIGKBS-JSAI, ICS-IPSJ, and IEICE-SIGAI on Active Mining)
,
PDFダウンロードページ PDFダウンロードページへ
抄録(和)
抄録(英) Constructing a classifier for structured data for which the attribute value type data representation is not straightforward is challenging. Graph is one such structured data and we developed DT-GBI in which B-GBI, a graph mining program, is recursively called to construct attributes in the form of subgraphs. B-GBI, although quite efficient, has some problem due to non backtracking greedy strategy with pairwise chunking. Overlapping subgraphs can't be searched and frequency counting of the same subgraph in a single graph has some ambiguity. Cl-GBI, recently proposed new approach that uses chunkingless chunks as pseud-nodes, solves this problem at the expense of computation cost. DT-C1GBI is an improved version of DT-GBI that uses Cl-GBI instead of B-GBI. Experimental results of DT-C1GBI are shown for hepatitis dataset with some discussion. The work is still in-progress.
キーワード(和)
キーワード(英)
資料番号 AI2004-26
発行日

研究会情報
研究会 AI
開催期間 2004/11/27(から1日開催)
開催地(和)
開催地(英)
テーマ(和)
テーマ(英)
委員長氏名(和)
委員長氏名(英)
副委員長氏名(和)
副委員長氏名(英)
幹事氏名(和)
幹事氏名(英)
幹事補佐氏名(和)
幹事補佐氏名(英)

講演論文情報詳細
申込み研究会 Artificial Intelligence and Knowledge-Based Processing (AI)
本文の言語 ENG
タイトル(和)
サブタイトル(和)
タイトル(英) Analysis of Hepatitis Dataset by Using Cl-GBI(Medical Active Mining)(Joint Workshop of Vietnamese Society of AI, SIGKBS-JSAI, ICS-IPSJ, and IEICE-SIGAI on Active Mining)
サブタイトル(和)
キーワード(1)(和/英)
第 1 著者 氏名(和/英) / AKIRA MOGI
第 1 著者 所属(和/英)
発表年月日 2004/11/27
資料番号 AI2004-26
巻番号(vol) vol.104
号番号(no) 485
ページ範囲 pp.-
ページ数 6
発行日