{"_buckets": {"deposit": "e2499f41-7389-4a8d-ab6d-cda0c2cfb4d4"}, "_deposit": {"created_by": 3, "id": "3457", "owners": [3], "pid": {"revision_id": 0, "type": "depid", "value": "3457"}, "status": "published"}, "_oai": {"id": "oai:repository.ninjal.ac.jp:00003457", "sets": ["508"]}, "author_link": ["11608", "11607"], "item_10002_biblio_info_40": {"attribute_name": "書誌情報", "attribute_value_mlt": [{"bibliographicIssueDates": {"bibliographicIssueDate": "2021-07", "bibliographicIssueDateType": "Issued"}, "bibliographicIssueNumber": "21", "bibliographicPageEnd": "132", "bibliographicPageStart": "123", "bibliographic_titles": [{"bibliographic_title": "国立国語研究所論集"}, {"bibliographic_title": "NINJAL Research Papers", "bibliographic_titleLang": "en"}]}]}, "item_10002_description_34": {"attribute_name": "著者所属", "attribute_value_mlt": [{"subitem_description": "神戸松蔭女子学院大学", "subitem_description_type": "Other"}]}, "item_10002_description_35": {"attribute_name": "著者所属(英)", "attribute_value_mlt": [{"subitem_description": "Kobe Shoin Women’s University", "subitem_description_type": "Other"}]}, "item_10002_description_36": {"attribute_name": "抄録", "attribute_value_mlt": [{"subitem_description": "旧字体と新字体の混在するテキストは，形態素解析において誤解析の原因となることが多く，その対策としては形態素解析辞書の記載に異体字を加える方法，そして予め漢字を新字体に置換しておく方法，また複数の辞書を使い分けるといった方法が考えられる。本稿では字体置換6通りと，辞書の使い分け3通りを掛け合わせた18組の組み合わせで國/国，會/会，關/関3対の旧/新字体の対を含んだテキストの形態素解析を行うことで，目的とする漢字を含む形態素がどれほど正確に切り出せるのかを検討した。データとして第1～10回までの国会会議録を用いた。結果は，漢字置換で隣接する漢字が旧字体の場合に旧字体に置換し，隣接しない場合は新字体とするという置換法（デフォルトを新字体とする日和見置換）と，すべてについて近代文語UniDicを用いるか，1949年の当用漢字字体表告示を境として，それ以前では近代文語UniDicを用い，それ以後では現代語書き言葉UniDicを用いる方法が，もっとも正確に当該漢字を含む短単位形態素を切り出せるというものであった。形態素解析辞書の記載に異体字を加える方法には，異体字が記載されていない形態素が出現した場合に対応ができないという欠点があるのに対して，漢字置換と辞書の使い分けを活用する方法は，そうした場合にも柔軟に対応が可能であるという利点があることを主張した。", "subitem_description_type": "Abstract"}]}, "item_10002_description_37": {"attribute_name": "抄録(英)", "attribute_value_mlt": [{"subitem_description": "Japanese texts containing a mixture of old-(kyūjitai) and new-(shinjitai) style kanji scripts pose a serious problem for an automatic morphological analyzer. However, recent developments in various dictionaries by era, undertaken by the corpora project at NINJAL, have brought about a new opportunity to solve this problem. Another promising solution is to replace the script in the text in some way, so that the analyzer can correctly identify the characters/morphemes. We designed an experiment with three dictionary selection methods and six replacement methods using three pairs of old/new kanji scripts (國/国, 會/会 and 關/関) to determine which combination would result in the most precise analysis. An analysis of the text data from the Minutes of the National Diet between 1947 and 1951 demonstrated that, of the 18 combinations, two dictionaries gave the best results. These were, the Contemporary Written Japanese UniDic dictionary up to the public notification of the Table of Script Styles of Jōyō Kanji on April 29, 1949, and The Modern Literary UniDic. With these, we coupled a replacement of a kanji script with an old counterpart when its immediate neighbor was also an old one, and with a new one when it was not. Although the addition of the different scripts to the dictionary entries would be another viable solution, our method is more desirable in that it is applicable to a wider range of texts without dictionary entry modifications.", "subitem_description_type": "Other"}]}, "item_10002_description_51": {"attribute_name": "フォーマット", "attribute_value_mlt": [{"subitem_description": "application/pdf", "subitem_description_type": "Other"}]}, "item_10002_identifier_registration": {"attribute_name": "ID登録", "attribute_value_mlt": [{"subitem_identifier_reg_text": "10.15084/00003440", "subitem_identifier_reg_type": "JaLC"}]}, "item_10002_publisher_39": {"attribute_name": "出版者", "attribute_value_mlt": [{"subitem_publisher": "国立国語研究所"}]}, "item_10002_source_id_41": {"attribute_name": "ISSN", "attribute_value_mlt": [{"subitem_source_identifier": "2186-1358", "subitem_source_identifier_type": "ISSN"}]}, "item_10002_title_29": {"attribute_name": "タイトル（ヨミ）", "attribute_value_mlt": [{"subitem_title": "シンカンジ ト キュウカンジ ガ コンザイ シタ テキスト カラ ノ タンタンイ ケイタイソ ノ チュウシュツ ニ ツイテ"}]}, "item_10002_version_type_52": {"attribute_name": "著者版フラグ", "attribute_value_mlt": [{"subitem_version_resource": "http://purl.org/coar/version/c_970fb48d4fbd8a85", "subitem_version_type": "VoR"}]}, "item_creator": {"attribute_name": "著者", "attribute_type": "creator", "attribute_value_mlt": [{"creatorNames": [{"creatorName": "松田, 謙次郎"}], "nameIdentifiers": [{"nameIdentifier": "11607", "nameIdentifierScheme": "WEKO"}]}, {"creatorNames": [{"creatorName": "MATSUDA, Kenjiro", "creatorNameLang": "en"}], "nameIdentifiers": [{"nameIdentifier": "11608", "nameIdentifierScheme": "WEKO"}]}]}, "item_files": {"attribute_name": "ファイル情報", "attribute_type": "file", "attribute_value_mlt": [{"accessrole": "open_date", "date": [{"dateType": "Available", "dateValue": "2021-07-17"}], "displaytype": "detail", "download_preview_message": "", "file_order": 0, "filename": "papers2108.pdf", "filesize": [{"value": "850.9 kB"}], "format": "application/pdf", "future_date_message": "", "is_thumbnail": false, "licensetype": "license_free", "mimetype": "application/pdf", "size": 850900.0, "url": {"label": "papers2108.pdf", "url": "https://repository.ninjal.ac.jp/record/3457/files/papers2108.pdf"}, "version_id": "1ddbe231-e48d-46f7-a6eb-a2cd2567613f"}]}, "item_keyword": {"attribute_name": "キーワード", "attribute_value_mlt": [{"subitem_subject": "形態素解析", "subitem_subject_scheme": "Other"}, {"subitem_subject": "国会会議録", "subitem_subject_scheme": "Other"}, {"subitem_subject": "旧字体", "subitem_subject_scheme": "Other"}, {"subitem_subject": "新字体", "subitem_subject_scheme": "Other"}, {"subitem_subject": "当用漢字字体表", "subitem_subject_scheme": "Other"}, {"subitem_subject": "morphological analysis", "subitem_subject_language": "en", "subitem_subject_scheme": "Other"}, {"subitem_subject": "Minutes of the National Diet", "subitem_subject_language": "en", "subitem_subject_scheme": "Other"}, {"subitem_subject": "kyūjitai", "subitem_subject_language": "en", "subitem_subject_scheme": "Other"}, {"subitem_subject": "shinjitai", "subitem_subject_language": "en", "subitem_subject_scheme": "Other"}, {"subitem_subject": "the Table of Script Styles of Tōyō Kanji", "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": "新漢字と旧漢字が混在したテキストからの短単位形態素の抽出について", "item_titles": {"attribute_name": "タイトル", "attribute_value_mlt": [{"subitem_title": "新漢字と旧漢字が混在したテキストからの短単位形態素の抽出について"}, {"subitem_title": "How to Correctly Morphologically Analyze Text Containing a Mixture of Old- and New-Style Kanji Scripts", "subitem_title_language": "en"}]}, "item_type_id": "10002", "owner": "3", "path": ["508"], "permalink_uri": "https://doi.org/10.15084/00003440", "pubdate": {"attribute_name": "公開日", "attribute_value": "2021-07-16"}, "publish_date": "2021-07-16", "publish_status": "0", "recid": "3457", "relation": {}, "relation_version_is_last": true, "title": ["新漢字と旧漢字が混在したテキストからの短単位形態素の抽出について"], "weko_shared_id": 3}