Soft-magin support vector machine maximizing geometric margins for multiclass classification

• Project/Area Number: 24500275
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Research Field: Sensitivity informatics/Soft computing
• Research Institution: Osaka University
• Principal Investigator: Tatsimi Keiji 大阪大学, 工学(系)研究科(研究院), 准教授 (30304017)
• Project Period (FY): 2012-04-01 – 2016-03-31
• Project Status: Completed (Fiscal Year 2015)
• Budget Amount: ¥5,070,000 (Direct Cost: ¥3,900,000、Indirect Cost: ¥1,170,000); Fiscal Year 2015: ¥1,170,000 (Direct Cost: ¥900,000、Indirect Cost: ¥270,000); Fiscal Year 2014: ¥1,300,000 (Direct Cost: ¥1,000,000、Indirect Cost: ¥300,000); Fiscal Year 2013: ¥1,300,000 (Direct Cost: ¥1,000,000、Indirect Cost: ¥300,000); Fiscal Year 2012: ¥1,300,000 (Direct Cost: ¥1,000,000、Indirect Cost: ¥300,000)
• Keywords: 多クラス識別問題 / 教師有り学習 / サポートベクトルマシン / 多目的最適化 / マージン最大化 / クラスタリング / 2次錐計画問題 / 多クラス識別 / 一対多手法 / 汎化性能 / 一括型手法 / 機械学習 / 一対一手法 / 多目的最適化問題

Outline of Final Research Achievements

We had proposed the multiobjective multiclass support vector machine (MMSVM), which can maximize the geometric margins for multiclass classification problem. In this study, we developed reduction methods of time and space computational complexity of the MMSVM, and extended the MMSVM into soft-margin models which can learn training data including outliers, where we compare the performances of various combinations from some restriction methods of the constraints of the MMSVM, several solving method of the multiobjective optimization, and two kinds of penalty functions. Among them, we verified that some propose methods, a data reduction method based on support vectors for the soft-margin MMSVM, an improved MMSVM based on the one-against-all method, an MMSVM based on the k-means method are more effective than existing methods.

Research Products

• [Journal Article] Support vector machines maximizing geometric margins for multi-class classification (2014)
  • Author(s): Keiji Tatsumi, Tetsuzo Tanino
  • Journal Title: TOP: An Official Journal of the Spanish Society of Statistics and Operations Research Volume: 22 Pages: 815-840
  • Peer Reviewed
• [Presentation] クラスタリングにより識別器候補を限定した多目的マルチクラス サポートベクトルマシンの計算量削減法 (2016)
  • Author(s): 巽啓司
  • Organizer: 計測自動制御学会 第43回知能システムシンポジウム
  • Place of Presentation: 室蘭工業大学（室蘭、北海道）
  • Year and Date: 2016-03-10
• [Presentation] A Complexity Reduction Method for the Multiobjective Multiclass Support Vector Machine (2014)
  • Author(s): Keiji Tatsumi, Tetsuzo Tanino
  • Organizer: The 20th Conference of the International Federation of Operational Research Societies
  • Place of Presentation: Centre de Convencions Internacional de Barcelona - CCIB, Barcelona, Spain
  • Year and Date: 2014-07-14
• [Presentation] 一対多手法に基づいた幾何マージン最大化手法のソフ トマージンモデルへの拡張 と学習時間削減 (2014)
  • Author(s): 巽啓司
  • Organizer: 日本オペレーションズ・リサーチ学会 2014年春季研究発表会
  • Place of Presentation: 大阪大学豊中キャンパス（大阪）
• [Presentation] 一対一手法に基づいた多クラス多目的サポートベ クトルマシンの学習 (2013)
  • Author(s): 巽啓司
  • Organizer: 第57回システム制御情報学会研究発表講演会
  • Place of Presentation: 兵庫県民会館（兵庫）
• [Presentation] 一対一手法に基づく多クラス多目的 サポートベクトルマシンに対する参 照点を利用した学習法 (2013)
  • Author(s): 巽啓司
  • Organizer: 日本オペレーションズ・リサーチ学会 2013年秋季研究発表会
  • Place of Presentation: 徳島大学 常三島キャンパス（徳島）
• [Presentation] A soft-margin multiobjective multiclass support vector machine with unified objective functions for maximizing geometric margins and minimizing slack variables (2012)
  • Author(s): 巽啓司
  • Organizer: 15th Czech-Japan Seminars on Data Analysis and Decision Making under Uncertainty
  • Place of Presentation: 大阪大学待兼山会館（大阪）
• [Presentation] Multiclass Support Vector Machines Based on Multiobjective Optimization (2012)
  • Author(s): 巽啓司
  • Organizer: 西安交通大学電子情報工学部
  • Place of Presentation: 西安交通大学電子情報工学部（西安・中華人民共和国）
  • Invited