Please use this identifier to cite or link to this item:

Title: Which Kinds of Trend Metrics Are More Effective for Emerging Trend Detection?
Other Titles: 科技前瞻趨勢預測方法的成效驗證
Authors: 曾元顯
Tseng, Yuen-Hsien
Hung, Wen-Chi;Lee, Yi-Yang
Keywords: 集群;趨勢指標;特徵趨勢;線性迴歸;評估;資訊檢索
;Trend metrics;Eigen-trend;Linear regression;Evaluation;Information retrieval
Date: 2009-03
Issue Date: 2016-06-24 17:10:27 (UTC+8)
Abstract: 科學計量學的科技前瞻方法常以各類參數觀察並預測趨勢,但是並未檢驗參數的有效性。本研究比較了數個趨勢觀察方法,利用資訊檢索評估相關排序的方法評估其排序效果,包括:趨勢呈現方式、趨勢公式以及時間區隔。以不同的領域、文件規模、以及主題來源進行成效比較。結果顯示時間序列線性迴歸斜率在各種情況下表現良好。本研究不僅提供科學計量學趨勢預測效果評估方法,對過去及未來的趨勢分析研究也提供了反思與洞察。
In scientometrics for trend analysis, parameter choices for observing trends are often made ad hoc in past studies. However, the effectiveness of these choices was hardly examined, quantitatively and comparatively. This work provides clues to better interpret the results when a certain parameter choice is made. Specifically, by sorting research topics in descending order of interest predicted by a trend metric and then by evaluating this ordering based on information retrieval measures, we compare a number of trend metrics (percentage of increase vs. regression slope), trend formulations (simple trend vs. eigen-trend), and options (various year spans and durations for prediction) in different domains (safety agriculture and information retrieval) with different collection scales (72,500 papers and 853 papers) to know which one leads to better trend observation. Our results show that the slope of linear regression on the time series performs constantly better than the others. More interestingly, this metric is robust under different conditions and is hardly affected even when the collection is split into arbitrary periods. Implications of these results are discussed. Our work not only provides a method to evaluate trend prediction performance for scientometrics, but also offers insights and reflections for past and future trend observation studies.
Relation: 圖書與資訊學刊, 68(1:1), 12-29
Journal of Librarianship and Information Science
Data Type: article
DOI 連結:
Appears in Collections:[圖資與檔案學刊] 期刊論文

Files in This Item:

File SizeFormat
68-2.pdf474KbAdobe PDF526View/Open

All items in 學術集成 are protected by copyright, with all rights reserved.

社群 sharing