{"created":"2023-05-15T14:42:37.189474+00:00","id":2961,"links":{},"metadata":{"_buckets":{"deposit":"3a4b1f44-fe5e-48c6-84d6-c93d8b3f4541"},"_deposit":{"created_by":3,"id":"2961","owners":[3],"pid":{"revision_id":0,"type":"depid","value":"2961"},"status":"published"},"_oai":{"id":"oai:ksu.repo.nii.ac.jp:00002961","sets":["14:226:229"]},"author_link":["7737"],"control_number":"2961","item_2_biblio_info_12":{"attribute_name":"書誌情報","attribute_value_mlt":[{"bibliographicIssueDates":{"bibliographicIssueDate":"2014-07","bibliographicIssueDateType":"Issued"},"bibliographicPageEnd":"86","bibliographicPageStart":"69","bibliographicVolumeNumber":"9","bibliographic_titles":[{"bibliographic_title":"京都産業大学総合学術研究所所報"}]}]},"item_2_description_10":{"attribute_name":"抄録(日)","attribute_value_mlt":[{"subitem_description":"これまでに，結合強度の値を従来の実数からファジィ数に拡張したニューラルネットと，その学習方法が提案されている．しかし，従来の学習方法は誤差逆伝播法を拡張した方法であったため，教師データが必要であった．一方，教師データを必要とせずにニューラルネットを学習させる方法として，進化的アルゴリズムを用いた強化学習方法が研究されており，ニューロエボリューションと呼ばれている．しかし，従来のニューロエボリューションにおいては，学習対象のニューラルネットは結合強度の値が実数である古典的なニューラルネットであり，前記のファジィニューラルネットの学習には適用されていなかった．これに対して本研究ではこれまでに，genotypeの値としてファジィ数を利用できるように進化的アルゴリズムを拡張する手法を提案した．この提案手法を用いれば，ファジィニューラルネットの進化的学習が可能になると考えられる．本稿では，ファジィニューラルネットの進化的学習に対する提案手法の有効性評価を行った．実験では，結合強度の値であるファジィ数を対称三角型とし，進化的アルゴリズムとしては遺伝的アルゴリズムを用いた．また，ファジィニューラルネットの学習目標は，実数の入力に対してファジィ数を出力する非線形な関数を近似することとした．実験の結果，教師データが与えられず学習目標の関数も明示的には与えられていないにも関わらず，提案手法によってファジィニューラルネットを自律的に進化させ，目標関数を近似させることができた．また，対称三角型ファジィ数を規定する方法として上限・下限モデルと中心・幅モデルを比較した結果，中心・幅モデルを用いたほうがより適合度の高いファジィニューラルネットが生成され，中心・幅モデルの利用が望ましいことがわかった．","subitem_description_type":"Other"}]},"item_2_description_11":{"attribute_name":"抄録(英)","attribute_value_mlt":[{"subitem_description":"The author has proposed an extension of genetic algorithm（GA）for solving fuzzy-valued optimization problems. In the proposed GA, genotype values are not real numbers but fuzzy numbers. Evolutionary processes in GA are extended so that GA can handle genotype instances with fuzzy numbers. In this article, the author applies the proposed method to evolving fuzzy neural networks（FNNs）. In the FNNs, values of weights and biases are not real numbers but fuzzy numbers. The ordinary GA cannot be applied to the neuroevolution of the FNNs because evolutionary processes in the ordinary GA are not designed to handle fuzzy numbers as genotype values. On the contrary, the proposed GA can adopt fuzzy numbers of FNN weights and biases directly as genotype values: the weights and biases are tuned by evolutionary operations of the proposed GA, not by the traditional back propagation algorithm. Experimental results showed that fuzzy neural networks evolved by the fuzzy GA could model hidden target fuzzy functions well despite the fact that no training data was explicitly provided. Besides, the experimental results revealed that, in the case of adopting symmetric triangular fuzzy numbers as fuzzy genotype values, the center and width model could contribute better than the lower and upper model did for evolving the FNNs.","subitem_description_type":"Other"}]},"item_2_description_15":{"attribute_name":"表示順","attribute_value_mlt":[{"subitem_description":"5","subitem_description_type":"Other"}]},"item_2_description_16":{"attribute_name":"アクセション番号","attribute_value_mlt":[{"subitem_description":"KJ00009319239","subitem_description_type":"Other"}]},"item_2_description_2":{"attribute_name":"ページ属性","attribute_value_mlt":[{"subitem_description":"P","subitem_description_type":"Other"}]},"item_2_description_8":{"attribute_name":"記事種別(日)","attribute_value_mlt":[{"subitem_description":"研究論文","subitem_description_type":"Other"}]},"item_2_description_9":{"attribute_name":"記事種別(英)","attribute_value_mlt":[{"subitem_description":"Paper","subitem_description_type":"Other"}]},"item_2_source_id_1":{"attribute_name":"雑誌書誌ID","attribute_value_mlt":[{"subitem_source_identifier":"AA11879037","subitem_source_identifier_type":"NCID"}]},"item_2_source_id_19":{"attribute_name":"ISSN","attribute_value_mlt":[{"subitem_source_identifier":"1348-8465","subitem_source_identifier_type":"PISSN"}]},"item_2_text_6":{"attribute_name":"著者所属(日)","attribute_value_mlt":[{"subitem_text_value":"京都産業大学コンピュータ理工学部"}]},"item_creator":{"attribute_name":"著者","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"岡田, 英彦","creatorNameLang":"ja"},{"creatorName":"OKADA, Hidehiko","creatorNameLang":"en"}],"nameIdentifiers":[{}]}]},"item_files":{"attribute_name":"ファイル情報","attribute_type":"file","attribute_value_mlt":[{"accessrole":"open_date","date":[{"dateType":"Available","dateValue":"2018-01-28"}],"displaytype":"detail","filename":"KJ00009319239.pdf","filesize":[{"value":"1.1 MB"}],"format":"application/pdf","licensetype":"license_note","mimetype":"application/pdf","url":{"label":"KJ00009319239.pdf","url":"https://ksu.repo.nii.ac.jp/record/2961/files/KJ00009319239.pdf"},"version_id":"ccfa5ae8-493c-458d-96d1-e9d8cfb6de77"}]},"item_keyword":{"attribute_name":"キーワード","attribute_value_mlt":[{"subitem_subject":"進化的アルゴリズム","subitem_subject_scheme":"Other"},{"subitem_subject":"evolutionary algorithms","subitem_subject_language":"en","subitem_subject_scheme":"Other"},{"subitem_subject":"fuzzy number","subitem_subject_language":"en","subitem_subject_scheme":"Other"},{"subitem_subject":"neural network","subitem_subject_language":"en","subitem_subject_scheme":"Other"},{"subitem_subject":"neuroevolution","subitem_subject_language":"en","subitem_subject_scheme":"Other"},{"subitem_subject":"遺伝的アルゴリズム","subitem_subject_scheme":"Other"},{"subitem_subject":"ニューラルネット","subitem_subject_scheme":"Other"},{"subitem_subject":"ファジィ数","subitem_subject_scheme":"Other"},{"subitem_subject":"ニューロエボリューション","subitem_subject_scheme":"Other"},{"subitem_subject":"genetic algorithm","subitem_subject_language":"en","subitem_subject_scheme":"Other"}]},"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":"ファジィ遺伝的アルゴリズムによるファジィニューラルネットの進化的学習","item_titles":{"attribute_name":"タイトル","attribute_value_mlt":[{"subitem_title":"ファジィ遺伝的アルゴリズムによるファジィニューラルネットの進化的学習","subitem_title_language":"ja"}]},"item_type_id":"2","owner":"3","path":["229"],"pubdate":{"attribute_name":"PubDate","attribute_value":"2018-01-28"},"publish_date":"2018-01-28","publish_status":"0","recid":"2961","relation_version_is_last":true,"title":["ファジィ遺伝的アルゴリズムによるファジィニューラルネットの進化的学習"],"weko_creator_id":"3","weko_shared_id":-1},"updated":"2023-09-04T06:37:27.650255+00:00"}