2003 Fiscal Year Annual Research Report

構造データからのアクティブマイニング

Research Project

Project/Area Number	13131206
Research Institution	Osaka University
Principal Investigator	元田浩大阪大学, 産業科学研究所, 教授 (00283804)
Co-Investigator(Kenkyū-buntansha)	HO Tu Bao 北陸先端科学技術大学院大学, 知識科学研究科, 教授 (60301199) 吉田哲也大阪大学, 産業科学研究所, 助手 (80294164) 鷲尾隆大阪大学, 産業科学研究所, 助教授 (00192815) 矢田勝俊関西大学, 商学部, 助教授 (00298811)
Keywords	機械学習 / 知識発見 / データマイニング / クラスタリング / 視覚化 / 時系列解析 / データ前処理
Research Abstract	研究実績は以下の通り. 1.専門家からのフィードバックに柔軟に対応できるように,データ準備に必要な一連の前処理プログラムをフィルタの集合として整備し,データ準備の効率化を図った. 2.グラフ構造データ用決定木構築手法DT-GBIにBeam探索と枝刈り機能を追加し,肝硬変患者の同定,肝炎タイプの同定,インターフェロン効果の予測に適用し,構築した決定木の予測精度を交差検定により評価し,改良効果を確認した. 3.3次元グラフマイニング手法を実際の化学構造データの解析に適用し,適切な原子間距離のレベル分けが重要であるとの知見を得た. 4.肝炎時系列データの特徴が長期的に変化するデータと短期的に変化するデータで違うことをデータから見いだし,それぞれに適した抽象化表現および典型的な変化パターンを,分類するためのアルゴリズムを開発し,獲得される膨大な量の規則から統計的有意性に基き,より信頼性の高い規則を抽出した. 5.不規則間隔時系列データの別の抽象化手法として,スケールが自動的に調整できるウェーブレット変換を提案し,性能を評価し,ユークリッド距離により時系列を処理する他のアルゴリズムと比較し同等以上の精度が得られることを確認した. 6.名辞属性は値同士の関係が直接定義できずデータ間の類似性の測定は困難である.属性の確率分布を用い、2つの名辞属性値間の非類似度をそれらの条件付確率分布間の非類似性の総計として,2つのデータ間の非類似度をこの各属性値間の非類似度の総計として評価する尺度を提案した. 7.XMLベースのデータマイニシグプラットフォームMUSASHIを開発し,オープンソースソフトウェアとして公開した.また,MUSASHI上で動くCRMシステムC-MUSASHIのプロトタイプを開発した.多くのボランティアに支えられるMUSASHIユーザー会が発足し,日本における普及体制が整いつつある.

Research Products

(18 results)

All Other

All Publications (18 results)

[Publications] A.Inokuchi: "Complete Mining of Frequent Patterns from Graphs : Mining Graph Data"Machine Learning. Vol.50, No.3. 321-354 (2003)
[Publications] 西村芳男: "AGMアルゴリズムの高速化と立体構造解析への適用"人工知能学会論文誌. Vol.18, No.5C. 257-268 (2003)
[Publications] S.Kawasaki: "Temporal Abstraction for Long-Term Changed Tasts in the Hepatitis Domain"Journal of Advanced Computational Intelligence & Intelligent Informatics. Vol.170, No.3. 348-354 (2003)
[Publications] K.Terai: "Business Process Semi-Automation Based on Business Model Management"International Journal of Intelligent Systems in Accounting, Finance and Management. Vol.11, No.4. 215-234 (2003)
[Publications] T.B.Ho: "A Knowledge Discovery System with Support for Model Selection and Visualization"Applied Intelligence, Kluwer Academic Publishers. Vol.19, No.1-2. 125-141 (2003)
[Publications] T.B.Ho: "Chance Discovery and Learning Minority Classes"Journal of New Generation Computing. Vol.21, No.2. 147-160 (2003)
[Publications] H.Liu: "Active Feature Selection Using Classes"Proceedings of the 7th Pacific-Asia Conference on Knowledge Discovery and Data Mining. LNAI2637. 474-485 (2003)
[Publications] W.Geamsakul: "Constructing a Decision Tree for Graph Structured Data"Proceedings of the First International Workshop on Mining Graphs, Trees and Sequences (MGTS-2003). 1-10 (2003)
[Publications] A.Inokuchi: "Specific Biasis for Mining Frequent Substructures"Proceedings of the First International Workshop on Mining Graphs, Trees and Sequences (MGTS-2003). 45-54 (2003)
[Publications] F.Adachi: "Development of Generic Search Method Based on Transformation Invariance"Proceedings of the 14th International Symposium on Methodologies for Intelligent Systems (ISMIS03). 486-495 (2003)
[Publications] T.D.Nguyen: "Discovery of Trends and States in Irregular Temporal Data Abstraction of Long-Term Changed Attributes in the Hepatitis Domain"Proceedings of the 6th International Conference on Discovery Science. 410-417 (2003)
[Publications] T.B.Ho: "Mining Hepatitis Data with Temporal Abstraction"Proceedings of ACM International Conference on Knowledge Discovery and Data Mining KDD-03. 369-377 (2003)
[Publications] S.Kawasaki: "Abstraction of Long-Term Changed Attributes in the Hepatitis Domain"Proceedings of the 7th International Conference on Knowledge-based Intelligent Information and Intelligent Informatics. 365-372 (2003)
[Publications] P.C.Nguyen: "Temporal Decomposition : A Promising Approach to VQ-based Speaker Identification"Proceedings of Intenational Fuzzy Systems Association World Congress. 184-187 (2003)
[Publications] K.Yada: "Knowledge Discovery Process from Sales Data"Proceedings of the 2003 IRMA International Conference. 161-187 (2003)
[Publications] T.Araki: "Chance Discovery from Consumer Research Using Key Graph"Proceedings of A Full-day Workshop in HCI International 2003. 45-54 (2003)
[Publications] N.Katoh: "Business Application for Sales Transaction Data by Using Genome Analyis Technology"Proceedings of the 6th International Conference of Discovery Science. 208-219 (2003)
[Publications] K.Yada: "Data Mining Oriented CRM Systems Based on MUSASHI : C-MUSASHI"Proceedings of the Second International Workshop on Active Mining. 52-61 (2003)

2003 Fiscal Year Annual Research Report

構造データからのアクティブマイニング

Principal Investigator

元田 浩 大阪大学, 産業科学研究所, 教授 (00283804)

Research Products

[Publications] A.Inokuchi: "Complete Mining of Frequent Patterns from Graphs : Mining Graph Data"Machine Learning. Vol.50, No.3. 321-354 (2003)

[Publications] 西村芳男: "AGMアルゴリズムの高速化と立体構造解析への適用"人工知能学会論文誌. Vol.18, No.5C. 257-268 (2003)

[Publications] S.Kawasaki: "Temporal Abstraction for Long-Term Changed Tasts in the Hepatitis Domain"Journal of Advanced Computational Intelligence & Intelligent Informatics. Vol.170, No.3. 348-354 (2003)

[Publications] K.Terai: "Business Process Semi-Automation Based on Business Model Management"International Journal of Intelligent Systems in Accounting, Finance and Management. Vol.11, No.4. 215-234 (2003)

[Publications] T.B.Ho: "A Knowledge Discovery System with Support for Model Selection and Visualization"Applied Intelligence, Kluwer Academic Publishers. Vol.19, No.1-2. 125-141 (2003)

[Publications] T.B.Ho: "Chance Discovery and Learning Minority Classes"Journal of New Generation Computing. Vol.21, No.2. 147-160 (2003)

[Publications] H.Liu: "Active Feature Selection Using Classes"Proceedings of the 7th Pacific-Asia Conference on Knowledge Discovery and Data Mining. LNAI2637. 474-485 (2003)

[Publications] W.Geamsakul: "Constructing a Decision Tree for Graph Structured Data"Proceedings of the First International Workshop on Mining Graphs, Trees and Sequences (MGTS-2003). 1-10 (2003)

[Publications] A.Inokuchi: "Specific Biasis for Mining Frequent Substructures"Proceedings of the First International Workshop on Mining Graphs, Trees and Sequences (MGTS-2003). 45-54 (2003)

[Publications] F.Adachi: "Development of Generic Search Method Based on Transformation Invariance"Proceedings of the 14th International Symposium on Methodologies for Intelligent Systems (ISMIS03). 486-495 (2003)

[Publications] T.D.Nguyen: "Discovery of Trends and States in Irregular Temporal Data Abstraction of Long-Term Changed Attributes in the Hepatitis Domain"Proceedings of the 6th International Conference on Discovery Science. 410-417 (2003)

[Publications] T.B.Ho: "Mining Hepatitis Data with Temporal Abstraction"Proceedings of ACM International Conference on Knowledge Discovery and Data Mining KDD-03. 369-377 (2003)

[Publications] S.Kawasaki: "Abstraction of Long-Term Changed Attributes in the Hepatitis Domain"Proceedings of the 7th International Conference on Knowledge-based Intelligent Information and Intelligent Informatics. 365-372 (2003)

[Publications] P.C.Nguyen: "Temporal Decomposition : A Promising Approach to VQ-based Speaker Identification"Proceedings of Intenational Fuzzy Systems Association World Congress. 184-187 (2003)

[Publications] K.Yada: "Knowledge Discovery Process from Sales Data"Proceedings of the 2003 IRMA International Conference. 161-187 (2003)

[Publications] T.Araki: "Chance Discovery from Consumer Research Using Key Graph"Proceedings of A Full-day Workshop in HCI International 2003. 45-54 (2003)

[Publications] N.Katoh: "Business Application for Sales Transaction Data by Using Genome Analyis Technology"Proceedings of the 6th International Conference of Discovery Science. 208-219 (2003)

[Publications] K.Yada: "Data Mining Oriented CRM Systems Based on MUSASHI : C-MUSASHI"Proceedings of the Second International Workshop on Active Mining. 52-61 (2003)

元田浩大阪大学, 産業科学研究所, 教授 (00283804)