Loading...
|
Please use this identifier to cite or link to this item:
https://nccur.lib.nccu.edu.tw/handle/140.119/53269
|
| Title: | 開發學術智慧、指標、影響力之模型與技術 - 以資訊檢索為例 |
| Other Titles: | &Quot;Developing Academic Intelligence, Index, and Impact Model and Technique Involving Information Retrieval&Quot |
| Authors: | 諶家蘭 |
| Contributors: | 國立政治大學會計學系
行政院國家科學委員會 |
| Keywords: | 管理;學術;資訊檢索 |
| Date: | 2011 |
| Issue Date: | 2012-06-26 14:58:20 (UTC+8) |
| Abstract: | 偵測新興研究議題對於學者專家而言是一個相當重要議題，學者專家如何以有限時 間和資源進行辨識和尋找同一和不同領域中新興研究議題，其重要性比投入已經成熟研 究議題進行研發，相對而言將帶來較大貢獻度和影響力。本研究試圖協助學者專家偵測 和辨識未來新興研究議題，從大量文獻中探究和預測學術智慧。本研究首先進行探索性 研究(the pilot study)，瞭解會議論文與期刊論文間，議題領導和跟隨之關聯性，並 建立本研究第一年和第二年之研究基礎。 本研究第一年試圖發展一套衡量工具以有效評估研究議題之生命發展周期，是否具 有學術價值？如何加以辨識和衡量？過去文獻中對於認定一個議題是否新興或成熟缺 乏一套有效和公允之衡量工具，本研究將應用文字探勘、資訊檢索、和演算法技術，結 合統計分析方法，建構一套檢測流程和指標，協助學者專家辨識和偵測論文中之領導趨 勢以發掘學術智慧，並且協助學者專家在發現和追蹤新興議題之過程中節省下大量時間 和人力。 本研究第二年試圖針對文獻中對於研究議題之新興和成熟度，僅著重於議題之出現 頻率而忽視議題之新穎度和潛力，過去文獻僅單一採用出現頻率曲線，僅能辨識已經成 熟議題及事後議題趨勢，造成落後指標，無法提供預測和認定。因此，本研究試圖發展 一套衡量新興議題潛力之方法，包括新穎度指標、發表量指標和偵測點指標，藉由這些 指標和曲線辨識和偵測出前導性議題趨勢，協助學者專家建構各自領域之新興議題指標 之生命周期。 本研究第二年將同時探討應用文字探勘、資訊檢索、和貝式估計方法，推估同一和 不同領域中相關學者專家和出版刊物對於該領域之重要性和影響力，預測未來具有潛力 之研究議題、學者專家及出版刊物，採用貝式估計方法試圖從已發表論文和被引用論文 之資訊，建構作者和刊物之事前機率和概似函數，計算出同一和不同領域之重要作者和 刊物之影響力，進而檢定這些議題是否成為新興議題。 本研究第一年和第二年之研究實驗設計將採用大型重要研究資料庫，分別進行第一 年和第二年之研究產出之驗證，實驗資料量將相當龐大完整，未來所得研究結果深具參 考價值。大型重要研究資料庫預計包括：ACM Digital Library、IEEE Computer Society、 Science Direct Onsite、ProQuest Databases，採用論文之Titles、Abstracts、Keywords 進 行文字探勘之實驗設計。
Detecting emerging topics is an important task for researchers. Researchers use their limited time and resources to examine emerging topics is relative brings more contribute and impact than the mature topics in the field. This two-year research project will present the endeavors that attempts by researchers to identify emerging topics and identify research intelligence via academic papers. This research attempts to identify the relationship between topics investigated by conference papers and journal papers as the pilot study that can help the research decrease the plenty of time and effort to detect all the academic papers. In the first year, this study will detect emerging topics and will develop a measurement tool based on lifecycle development that can assess the research potential of the topics owing to the lack of an effective measurement tool for assessing whether the research topic is emerging and worth to put in more effort. In the second year, this study will apply the Bayesian estimation approach to estimate the possible impact of authors and publications on a topic and to identify candidate emerging topics via the combination of the influential authors and publications. In the pilot study, this research selects thousands of papers in data mining and information retrieval from well-known databases and shows that the topics covered by conference papers in a year often leads to similar topics covered by journal papers in the subsequent year and vice versa. The past literature assessing the maturity of the topic focused solely on frequency while ignoring the novelty of the topics. On the other hand, some existing researches only detected the new topic and disregarded its future research potential. Furthermore, analyzing topic importance based on their published frequency means topics will only be deemed important after they are already mature, resulting in an index that only looks backwards. Therefore, in the first year, this research attempts to develop indices for resolving this dilemma, and also attempts to develop a method of assessing the degree to which the topic is emerging. These indices include the novelty index, published volume index and detection point. Using the indices and curves as a basis can yield more forward information in emerging topic detection works and help researchers construct emerging topics detecting indices in their own fields. In the second year, the research adopts the concept that topics presented by authors and publications that have more influential topics presented by the author and publications will be likely to become emerging topics. Consequently, this investigation applies the Bayesian estimation approach and attempts to use the published information and citation analysis to construct the estimation function of the prior probability and likelihood function and calculate the influence of authors and publications. Papers published by authors working within their own field and the publication which is impact of the field will attract greater attention and be viewed as more worthwhile, and thus will be more likely to become emerging topics. The research aims to help researchers devise their own field detecting criterion detecting emerging topics and author-publication power. The research experiments will be conducted in the use of the main research databases including the ACM Digital Library, the IEEE Computer Society, the Science Direct Onsite, and the ProQuest Databases. The research experiments will be large scale both in the first year study and the second year study. Four main descriptors will be adopted to undertake the text mining and information retrieval in the first year’s and the second year’s researches. They are the titles, abstracts, keywords, and full text. |
| Relation: | 基礎研究
學術補助
研究期間:10008~ 10107
研究經費:907仟元 |
| Data Type: | report |
| Appears in Collections: | [會計學系] 國科會研究計畫
|
Files in This Item:
| File |
Description |
Size | Format | |
|---|
| 99-2410-H-004-107-MY2.pdf | | 1856Kb | Adobe PDF2 | 1111 | View/Open |
|
All items in 政大典藏 are protected by copyright, with all rights reserved.
|
