{"created":"2023-05-15T14:21:33.024282+00:00","id":3242,"links":{},"metadata":{"_buckets":{"deposit":"b4792223-8803-4c94-bd67-514f6924a4c1"},"_deposit":{"created_by":3,"id":"3242","owners":[3],"pid":{"revision_id":0,"type":"depid","value":"3242"},"status":"published"},"_oai":{"id":"oai:bunkyo.repo.nii.ac.jp:00003242","sets":["1:26:230"]},"author_link":["4197"],"item_5_biblio_info_13":{"attribute_name":"書誌情報","attribute_value_mlt":[{"bibliographicIssueDates":{"bibliographicIssueDate":"2001-01-01"},"bibliographicPageEnd":"62","bibliographicPageStart":"1","bibliographicVolumeNumber":"26","bibliographic_titles":[{"bibliographic_title":"情報研究"},{"bibliographic_title":"Information and Communication Studies"}]}]},"item_5_date_43":{"attribute_name":"作成日","attribute_value_mlt":[{"subitem_date_issued_datetime":"2011-02-23"}]},"item_5_description_12":{"attribute_name":"抄録","attribute_value_mlt":[{"subitem_description":"　SS理論と名付けられたパターン認識の数学的理論に登場するRECOGNITRONは、処理の対象とする問題の入力パターンψに対応し、\"axiom 1を満たすパターンモデル\"Tφを求め、Tφを恰も、φかのように扱う。このとき、写像Tはパターンモデル構成作用素と呼ばれる。axiom 2, 3を各々満たす類似度関数SM,大分類関数BSCを構成すれば、RECOGNITRONはφに関する連想形認識方程式を解くことによって、φから連想されるパターンと、φの帰属するカテゴリを求めることができる。\n　本論文では、あるカテゴリに帰属するか否かに分類される訓練データに関し、2分割された訓練データ間のマージンが最大になるような超平面を求める2カテゴリ学習分類器、サポートベクタマシン(SVM)の理論を適用し、axiom 3を満たす大分類関数BSCを設計する手法が提案される。\n　計算論的学習理論の1つとしての\"適応的ブースティングアルゴリズムAda Boost\"を適用して、BSCを設計できることは既に示されている。完全に線形分離でなくても分類誤差を考慮に入れて分離境界を与える超平面を決定するSVM理論の適用により、BSCの設計が訓練パターン集合について適切に設計できる1つの手法が得らている。\n\\n　RECOGNITRON appearing in a mathematical theory of recognizing patterns named SS-theory seeks from an input original pattern φ in question to be recognized a corresponding pattern-model Tφ which must satisfy axiom 1 suggested by S.Suzuki, and treats Tφ as though Tφ would be φ. The mapping T is called a model-construction operator. Provided that a similarity-measure function SM and a rough classifier BSC are constructed so as to respectively satisfy axiom 2 and axiom RECOGNITRON can determine a pattern recalled from pattern φ and a category to which φ belongs by solving an associative equation of recognition about φ.\n　In this paper, a BSC is designed according to a theory of support-vector machine(SVM). A learning machine SVM which can divide into two subsets of patterns seeks for two hyperplanes whose margin is maximized for a training set of patterns.\n　It was evident that BSC could be designed by applying. a boosting algorithm Ada Boost in a computational learning theory. SVM has an ablity of determining two hyperplanes which give two boundaries considering an error of classification whatever the traing set may not be linearly separable. Therefore a method of designing BSC is obtained with the object of adjusting RECOGNITRON to the training set.","subitem_description_type":"Abstract"}]},"item_5_description_38":{"attribute_name":"フォーマット","attribute_value_mlt":[{"subitem_description":"application/pdf","subitem_description_type":"Other"}]},"item_5_source_id_19":{"attribute_name":"ISSN","attribute_value_mlt":[{"subitem_source_identifier":"03893367"}]},"item_5_text_39":{"attribute_name":"本文言語","attribute_value_mlt":[{"subitem_text_value":"日本語"}]},"item_5_text_42":{"attribute_name":"ID","attribute_value_mlt":[{"subitem_text_value":"BKSJ260002"}]},"item_5_text_7":{"attribute_name":"Author","attribute_value_mlt":[{"subitem_text_value":"Suzuki, Shoichi"}]},"item_5_text_8":{"attribute_name":"所属機関","attribute_value_mlt":[{"subitem_text_value":"文教大学情報学部"}]},"item_5_text_9":{"attribute_name":"Institution","attribute_value_mlt":[{"subitem_text_value":"Bunkyo University Faculty of Information and Communications"}]},"item_5_version_type_35":{"attribute_name":"著者版フラグ","attribute_value_mlt":[{"subitem_version_type":"VoR"}]},"item_creator":{"attribute_name":"著者","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"鈴木, 昇一"}],"nameIdentifiers":[{}]}]},"item_files":{"attribute_name":"ファイル情報","attribute_type":"file","attribute_value_mlt":[{"accessrole":"open_date","date":[{"dateType":"Available","dateValue":"2018-03-24"}],"displaytype":"detail","filename":"BKSJ260002.pdf","filesize":[{"value":"3.2 MB"}],"format":"application/pdf","licensetype":"license_note","mimetype":"application/pdf","url":{"label":"BKSJ260002.pdf","url":"https://bunkyo.repo.nii.ac.jp/record/3242/files/BKSJ260002.pdf"},"version_id":"3435adc4-5e35-4b70-ae0c-60c8016a2cc8"}]},"item_keyword":{"attribute_name":"キーワード","attribute_value_mlt":[{"subitem_subject":"パターン認識の数学的理論(SS理論)"},{"subitem_subject":"モデル構成作用素"},{"subitem_subject":"類似度関数"},{"subitem_subject":"大分類関数"},{"subitem_subject":"2次計画問題"},{"subitem_subject":"サポートベクタ"}]},"item_language":{"attribute_name":"言語","attribute_value_mlt":[{"subitem_language":"jpn"}]},"item_resource_type":{"attribute_name":"資源タイプ","attribute_value_mlt":[{"resourcetype":"departmental bulletin paper","resourceuri":"http://purl.org/coar/resource_type/c_6501"}]},"item_title":"Support Vector Machineを利用した大分類関数の構成","item_titles":{"attribute_name":"タイトル","attribute_value_mlt":[{"subitem_title":"Support Vector Machineを利用した大分類関数の構成"},{"subitem_title":"A Construction of a Rough Classifier Having a Support Vector Machine as Its Structure"}]},"item_type_id":"5","owner":"3","path":["230"],"pubdate":{"attribute_name":"公開日","attribute_value":"2011-02-23"},"publish_date":"2011-02-23","publish_status":"0","recid":"3242","relation_version_is_last":true,"title":["Support Vector Machineを利用した大分類関数の構成"],"weko_creator_id":"3","weko_shared_id":-1},"updated":"2023-05-16T16:45:20.002325+00:00"}