{"created":"2023-03-30T09:23:56.430332+00:00","id":3043,"links":{},"metadata":{"_buckets":{"deposit":"f2fa535c-3afa-4644-b2bc-216942a07d83"},"_deposit":{"created_by":7,"id":"3043","owners":[7],"pid":{"revision_id":0,"type":"depid","value":"3043"},"status":"published"},"_oai":{"id":"oai:nied-repo.bosai.go.jp:00003043","sets":[]},"author_link":[],"item_10001_biblio_info_7":{"attribute_name":"書誌情報","attribute_value_mlt":[{"bibliographicIssueDates":{"bibliographicIssueDate":"2020-06","bibliographicIssueDateType":"Issued"},"bibliographicPageStart":"4Rin194","bibliographicVolumeNumber":"JSAI2020","bibliographic_titles":[{"bibliographic_title":"人工知能学会全国大会論文集","bibliographic_titleLang":"ja"},{"bibliographic_title":"Proceedings of the Annual Conference of JSAI","bibliographic_titleLang":"en"}]}]},"item_10001_description_5":{"attribute_name":"抄録","attribute_value_mlt":[{"subitem_description":"機械学習のブラックボックス性に関しては多くの議論があり、防災や医療、金融などの分野において人工知能技術を活用していく際には人工知能に説明や解釈が求められるケースが出てくると想定される。そのため、説明可能AIに向けた取り組みが今後ますます重要になっていくと考えられる。本研究では、地震動指標のランダムフォレスト予測器を対象として、その予測結果および機械学習モデル自体の説明・解釈を試みる。説明変数の重要度の評価や機械学習モデルにおける説明変数と目的変数の関係に関する議論を通じて、予測対象となる地表最大加速とへの震央距離やMwの寄与に関してはこれまでの地震学の知見と整合する一方で、震源深さやD1400，Vs30の寄与に関してはこれまで考えられてきたものよりも複雑なモデルを機械学習モデルは示唆していることが分かった。また弱学習器の予測を出力してアンサンブル学習による予測の中身を調べることで、学習データの不均衡が機械学習モデルに強い影響を与えていることもわかった。","subitem_description_language":"ja","subitem_description_type":"Other"},{"subitem_description":"We try to explain and interpret the random-forest predictor of earthquake ground-motion intensity and its prediction results. This study demonstrates that although the relationship of ground-motion intensity with earthquake magnitude or epicentral distance in the machine-learning model is consistent with the knowledge of seismology, the effects of source depth and site condition on ground-motion intensity in the machine-learning model are more complex than the assumptions of previous studies. The visualization of weak learners of the random forest indicates that its prediction is largely affected by the biased distribution of training data-set. However, the data-set is very biased and there are few strong ground-motion records. This causes the underestimation of the predictor for strong ground-motions. To overcome this problem, in this study, we suggest two approaches: one is the weighting of train data, and the other is the hybrid method integrating the conventional ground motion prediction equation and a machine learning approach. The verification using test data indicates that the hybrid method can largely improve the underestimation, although the underestimation still remains in predicting very strong groundmotions (>1000 gal).","subitem_description_language":"en","subitem_description_type":"Other"}]},"item_10001_publisher_8":{"attribute_name":"出版者","attribute_value_mlt":[{"subitem_publisher":"一般社団法人 人工知能学会","subitem_publisher_language":"ja"},{"subitem_publisher":"The Japanese Society for Artificial Intelligence","subitem_publisher_language":"en"}]},"item_10001_relation_14":{"attribute_name":"DOI","attribute_value_mlt":[{"subitem_relation_type_id":{"subitem_relation_type_id_text":"10.11517/pjsai.JSAI2020.0_4Rin194"}}]},"item_creator":{"attribute_name":"著者","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"久保久彦","creatorNameLang":"ja"},{"creatorName":"Hisahiko Kubo","creatorNameLang":"en"}]},{"creatorNames":[{"creatorName":"功刀卓","creatorNameLang":"ja"},{"creatorName":"Takashi Kunugi","creatorNameLang":"en"}]},{"creatorNames":[{"creatorName":"鈴木亘","creatorNameLang":"ja"},{"creatorName":"Wataru Suzuki","creatorNameLang":"en"}]},{"creatorNames":[{"creatorName":"木村武志","creatorNameLang":"ja"},{"creatorName":"Takeshi Kimura","creatorNameLang":"en"}]},{"creatorNames":[{"creatorName":"青井真","creatorNameLang":"ja"},{"creatorName":"Shin Aoi","creatorNameLang":"en"}]}]},"item_language":{"attribute_name":"言語","attribute_value_mlt":[{"subitem_language":"jpn"}]},"item_title":"説明可能な地震動指標のランダムフォレスト予測器の構築に向けた取り組み","item_titles":{"attribute_name":"タイトル","attribute_value_mlt":[{"subitem_title":"説明可能な地震動指標のランダムフォレスト予測器の構築に向けた取り組み","subitem_title_language":"ja"},{"subitem_title":"Construction of explainable random forest predictor for ground-motion intensity","subitem_title_language":"en"}]},"item_type_id":"40001","owner":"7","path":["1670839190650"],"pubdate":{"attribute_name":"PubDate","attribute_value":"2023-09-20"},"publish_date":"2023-09-20","publish_status":"0","recid":"3043","relation_version_is_last":true,"title":["説明可能な地震動指標のランダムフォレスト予測器の構築に向けた取り組み"],"weko_creator_id":"7","weko_shared_id":-1},"updated":"2023-09-20T08:10:33.366103+00:00"}