WEKO3
アイテム
{"_buckets": {"deposit": "5c90670b-2eb6-4cf5-9b99-a8cf8cb74e4e"}, "_deposit": {"created_by": 21, "id": "1682", "owners": [21], "pid": {"revision_id": 0, "type": "depid", "value": "1682"}, "status": "published"}, "_oai": {"id": "oai:ir.soken.ac.jp:00001682", "sets": ["19"]}, "author_link": ["0", "0", "0"], "item_1_biblio_info_21": {"attribute_name": "書誌情報(ソート用)", "attribute_value_mlt": [{"bibliographicIssueDates": {"bibliographicIssueDate": "2010-03-24", "bibliographicIssueDateType": "Issued"}, "bibliographic_titles": [{}]}]}, "item_1_creator_2": {"attribute_name": "著者名", "attribute_type": "creator", "attribute_value_mlt": [{"creatorNames": [{"creatorName": "CHANLEKHA, HUTCHATAI"}], "nameIdentifiers": [{"nameIdentifier": "0", "nameIdentifierScheme": "WEKO"}]}]}, "item_1_creator_3": {"attribute_name": "フリガナ", "attribute_type": "creator", "attribute_value_mlt": [{"creatorNames": [{"creatorName": "チャンレカ, ハッチャタイ"}], "nameIdentifiers": [{"nameIdentifier": "0", "nameIdentifierScheme": "WEKO"}]}]}, "item_1_date_granted_11": {"attribute_name": "学位授与年月日", "attribute_value_mlt": [{"subitem_dategranted": "2010-03-24"}]}, "item_1_degree_grantor_5": {"attribute_name": "学位授与機関", "attribute_value_mlt": [{"subitem_degreegrantor": [{"subitem_degreegrantor_name": "総合研究大学院大学"}]}]}, "item_1_degree_name_6": {"attribute_name": "学位名", "attribute_value_mlt": [{"subitem_degreename": "博士(情報学)"}]}, "item_1_description_1": {"attribute_name": "ID", "attribute_value_mlt": [{"subitem_description": "2010034", "subitem_description_type": "Other"}]}, "item_1_description_12": {"attribute_name": "要旨", "attribute_value_mlt": [{"subitem_description": "Public concern over the spread of infectious diseases such as avian H5N1 influenza and swine flu (H1N1) influenza A has underscored the importance of health surveillance\r\nsystems for the speedy and precise detection of disease outbreaks. However, two key barriers faced by the current web-based health surveillance systems are their inability to\u003cbr /\u003e(a) understand complex geo-temporal attributes of events and (b) to obtain the levels of\u003cbr /\u003egeo-temporal recognition. In this thesis, I develop a novel framework as a means to\u003cbr /\u003eovercome these limitations. This framework is called spatiotemporal zoning.\u003cbr /\u003e The objective of the spatiotemporal zoning scheme is to enable language echnology\u003cbr /\u003esoftware to partition text into segments based on the spatiotemporal characteristics of its\u003cbr /\u003econtent. Each segment, which is called a text zone, contains a set of events that occurred\u003cbr /\u003eat the same geographical location in the same time frame. The capability of associating\u003cbr /\u003eevents reported in each text segment with the most specific spatial and temporal\u003cbr /\u003einformation available in news reports enables simple techniques to be employed for\u003cbr /\u003edetecting specific outbreak locations. These techniques could be, for example, text\u003cbr /\u003eclassification to detect text segments that indicate outbreak situations. At the same time,\u003cbr /\u003efalse alarms about past outbreaks can be avoided by taking the temporal information\u003cbr /\u003eabout the events into consideration.\u003cbr /\u003e I created a representative corpus in order to demonstrate that spatiotemporal zoning\u003cbr /\u003ecan be automatically and manually applied to unrestricted text. The corpus consisted of\u003cbr /\u003e100 news articles from multiple news agencies reporting on various disease outbreaks in different parts of the world.\u003cbr /\u003e To study the reliability of spatiotemporal zoning, an experiment was conducted in which three annotators were recruited to annotate the same set of documents according to the annotation guidelines and the agreement between these annotators was then analyzed. Several statistical measures, namely kappa, Krippendorff’s alpha (α), and the percentage agreement, were used for quantitatively measuring the agreement. The results showed that the level of agreement kappa was more than 0.9 on average for event type and temporal attribute annotations, and it was only a slight lower for annotating spatial attributes.\u003cbr /\u003e The task of spatiotemporal zoning can be separated into 3 main steps. (1)Document pre-processing: This step provides the basic elements for zone attribute analysis and was done automatically using natural language processing software . (2) Zone attribute annotation: Each event-predicate is analyzed to recognize its class, spatial and temporal attributes. (3) Zone boundary generation: This step is done based on the attribute values of each event-predicate. For spatiotemporal zone annotation, the study of automatic zone attribute annotation was done for each group of zone attributes, i.e., event type recognition, temporal attributes recognition, and spatial attribute recognition.\u003cbr /\u003e To automatically classify event expressions, i.e. zone type recognition, Conditional Random Fields (CRFs) was used to incorporate various sets of text features into a classifier.\u003cbr /\u003e To recognize spatial information, several approaches, ranging from simple techniques such as the commonly used heuristic-based approach to the more sophisticated achine learning approach, were experimented. I also explored various feature sets and feature encoding strategies in order to determine the best ones for recognizing spatial attributes.\u003cbr /\u003e For temporal attribute recognition, I took a rule-based approach to recognizing an event\u0027s temporal information. However, one of the problems is that in many cases the same event is repeatedly mentioned whereas the time of its occurrence is stated only once. To improve the system\u0027s ability to recognize the temporal information, I employed a simple heuristic that helps to identify linguistic expressions referring to the same events.\u003cbr /\u003e The above studies that I undertook prove that spatiotemporal zoning is reliable. Moreover, the results from automatic zone attribute recognition show that this scheme can be done automatically with a reliable level of performance.", "subitem_description_type": "Other"}]}, "item_1_description_18": {"attribute_name": "フォーマット", "attribute_value_mlt": [{"subitem_description": "application/pdf", "subitem_description_type": "Other"}]}, "item_1_description_7": {"attribute_name": "学位記番号", "attribute_value_mlt": [{"subitem_description": "総研大甲第1338号", "subitem_description_type": "Other"}]}, "item_1_select_14": {"attribute_name": "所蔵", "attribute_value_mlt": [{"subitem_select_item": "有"}]}, "item_1_select_16": {"attribute_name": "複写", "attribute_value_mlt": [{"subitem_select_item": "印刷物から複写可"}]}, "item_1_select_17": {"attribute_name": "公開状況", "attribute_value_mlt": [{"subitem_select_item": "全文公開可"}]}, "item_1_select_8": {"attribute_name": "研究科", "attribute_value_mlt": [{"subitem_select_item": "複合科学研究科"}]}, "item_1_select_9": {"attribute_name": "専攻", "attribute_value_mlt": [{"subitem_select_item": "17 情報学専攻"}]}, "item_1_text_10": {"attribute_name": "学位授与年度", "attribute_value_mlt": [{"subitem_text_value": "2009"}]}, "item_creator": {"attribute_name": "著者", "attribute_type": "creator", "attribute_value_mlt": [{"creatorNames": [{"creatorName": "CHANLEKHA, Hutchatai", "creatorNameLang": "en"}], "nameIdentifiers": [{"nameIdentifier": "0", "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": "2016-02-17"}], "displaytype": "simple", "download_preview_message": "", "file_order": 0, "filename": "甲1338_要旨.pdf", "filesize": [{"value": "354.3 kB"}], "format": "application/pdf", "future_date_message": "", "is_thumbnail": false, "licensetype": "license_11", "mimetype": "application/pdf", "size": 354300.0, "url": {"label": "要旨・審査要旨", "url": "https://ir.soken.ac.jp/record/1682/files/甲1338_要旨.pdf"}, "version_id": "a6420545-5e08-437a-8471-f6e031841377"}, {"accessrole": "open_date", "date": [{"dateType": "Available", "dateValue": "2016-02-17"}], "displaytype": "simple", "download_preview_message": "", "file_order": 1, "filename": "甲1338_本文.pdf", "filesize": [{"value": "1.3 MB"}], "format": "application/pdf", "future_date_message": "", "is_thumbnail": false, "licensetype": "license_11", "mimetype": "application/pdf", "size": 1300000.0, "url": {"label": "本文", "url": "https://ir.soken.ac.jp/record/1682/files/甲1338_本文.pdf"}, "version_id": "7f1d0ca8-c92d-4388-8c27-19e9590471ac"}]}, "item_language": {"attribute_name": "言語", "attribute_value_mlt": [{"subitem_language": "eng"}]}, "item_resource_type": {"attribute_name": "資源タイプ", "attribute_value_mlt": [{"resourcetype": "thesis", "resourceuri": "http://purl.org/coar/resource_type/c_46ec"}]}, "item_title": "Document Zoning for Enhancing Spatial and Temporal Understanding in Web-based Health Surveillance Systems", "item_titles": {"attribute_name": "タイトル", "attribute_value_mlt": [{"subitem_title": "Document Zoning for Enhancing Spatial and Temporal Understanding in Web-based Health Surveillance Systems"}, {"subitem_title": "Document Zoning for Enhancing Spatial and Temporal Understanding in Web-based Health Surveillance Systems", "subitem_title_language": "en"}]}, "item_type_id": "1", "owner": "21", "path": ["19"], "permalink_uri": "https://ir.soken.ac.jp/records/1682", "pubdate": {"attribute_name": "公開日", "attribute_value": "2011-01-18"}, "publish_date": "2011-01-18", "publish_status": "0", "recid": "1682", "relation": {}, "relation_version_is_last": true, "title": ["Document Zoning for Enhancing Spatial and Temporal Understanding in Web-based Health Surveillance Systems"], "weko_shared_id": -1}
Document Zoning for Enhancing Spatial and Temporal Understanding in Web-based Health Surveillance Systems
https://ir.soken.ac.jp/records/1682
https://ir.soken.ac.jp/records/16825e85a76f-2d0b-41ef-9414-97115408d44e
名前 / ファイル | ライセンス | アクション |
---|---|---|
![]() |
||
![]() |
Item type | 学位論文 / Thesis or Dissertation(1) | |||||
---|---|---|---|---|---|---|
公開日 | 2011-01-18 | |||||
タイトル | ||||||
タイトル | Document Zoning for Enhancing Spatial and Temporal Understanding in Web-based Health Surveillance Systems | |||||
タイトル | ||||||
言語 | en | |||||
タイトル | Document Zoning for Enhancing Spatial and Temporal Understanding in Web-based Health Surveillance Systems | |||||
言語 | ||||||
言語 | eng | |||||
資源タイプ | ||||||
資源タイプ識別子 | http://purl.org/coar/resource_type/c_46ec | |||||
資源タイプ | thesis | |||||
著者名 |
CHANLEKHA, HUTCHATAI
× CHANLEKHA, HUTCHATAI |
|||||
フリガナ |
チャンレカ, ハッチャタイ
× チャンレカ, ハッチャタイ |
|||||
著者 |
CHANLEKHA, Hutchatai
× CHANLEKHA, Hutchatai |
|||||
学位授与機関 | ||||||
学位授与機関名 | 総合研究大学院大学 | |||||
学位名 | ||||||
学位名 | 博士(情報学) | |||||
学位記番号 | ||||||
内容記述タイプ | Other | |||||
内容記述 | 総研大甲第1338号 | |||||
研究科 | ||||||
値 | 複合科学研究科 | |||||
専攻 | ||||||
値 | 17 情報学専攻 | |||||
学位授与年月日 | ||||||
学位授与年月日 | 2010-03-24 | |||||
学位授与年度 | ||||||
2009 | ||||||
要旨 | ||||||
内容記述タイプ | Other | |||||
内容記述 | Public concern over the spread of infectious diseases such as avian H5N1 influenza and swine flu (H1N1) influenza A has underscored the importance of health surveillance systems for the speedy and precise detection of disease outbreaks. However, two key barriers faced by the current web-based health surveillance systems are their inability to<br />(a) understand complex geo-temporal attributes of events and (b) to obtain the levels of<br />geo-temporal recognition. In this thesis, I develop a novel framework as a means to<br />overcome these limitations. This framework is called spatiotemporal zoning.<br /> The objective of the spatiotemporal zoning scheme is to enable language echnology<br />software to partition text into segments based on the spatiotemporal characteristics of its<br />content. Each segment, which is called a text zone, contains a set of events that occurred<br />at the same geographical location in the same time frame. The capability of associating<br />events reported in each text segment with the most specific spatial and temporal<br />information available in news reports enables simple techniques to be employed for<br />detecting specific outbreak locations. These techniques could be, for example, text<br />classification to detect text segments that indicate outbreak situations. At the same time,<br />false alarms about past outbreaks can be avoided by taking the temporal information<br />about the events into consideration.<br /> I created a representative corpus in order to demonstrate that spatiotemporal zoning<br />can be automatically and manually applied to unrestricted text. The corpus consisted of<br />100 news articles from multiple news agencies reporting on various disease outbreaks in different parts of the world.<br /> To study the reliability of spatiotemporal zoning, an experiment was conducted in which three annotators were recruited to annotate the same set of documents according to the annotation guidelines and the agreement between these annotators was then analyzed. Several statistical measures, namely kappa, Krippendorff’s alpha (α), and the percentage agreement, were used for quantitatively measuring the agreement. The results showed that the level of agreement kappa was more than 0.9 on average for event type and temporal attribute annotations, and it was only a slight lower for annotating spatial attributes.<br /> The task of spatiotemporal zoning can be separated into 3 main steps. (1)Document pre-processing: This step provides the basic elements for zone attribute analysis and was done automatically using natural language processing software . (2) Zone attribute annotation: Each event-predicate is analyzed to recognize its class, spatial and temporal attributes. (3) Zone boundary generation: This step is done based on the attribute values of each event-predicate. For spatiotemporal zone annotation, the study of automatic zone attribute annotation was done for each group of zone attributes, i.e., event type recognition, temporal attributes recognition, and spatial attribute recognition.<br /> To automatically classify event expressions, i.e. zone type recognition, Conditional Random Fields (CRFs) was used to incorporate various sets of text features into a classifier.<br /> To recognize spatial information, several approaches, ranging from simple techniques such as the commonly used heuristic-based approach to the more sophisticated achine learning approach, were experimented. I also explored various feature sets and feature encoding strategies in order to determine the best ones for recognizing spatial attributes.<br /> For temporal attribute recognition, I took a rule-based approach to recognizing an event's temporal information. However, one of the problems is that in many cases the same event is repeatedly mentioned whereas the time of its occurrence is stated only once. To improve the system's ability to recognize the temporal information, I employed a simple heuristic that helps to identify linguistic expressions referring to the same events.<br /> The above studies that I undertook prove that spatiotemporal zoning is reliable. Moreover, the results from automatic zone attribute recognition show that this scheme can be done automatically with a reliable level of performance. |
|||||
所蔵 | ||||||
値 | 有 | |||||
フォーマット | ||||||
内容記述タイプ | Other | |||||
内容記述 | application/pdf |