非結構性資料影響力評估處理流程：以施政方針及施政報告為例

林峰正 ； 黃碧慧 ； 陳昶憲 ； 陳建元

《TANET2017 臺灣網際網路研討會》 2017年 (2017/10) Pp. 1483-1488

https://doi.org/10.6728/TANET.201710.0260

摘要

近年民眾參與意識抬頭，開放資料的出現，民眾可共同參與管理、決策，且資訊的流動與公開增加了政策的推動速度，而網際網路的快速發展，加上資訊爆炸的年代，使得資料檢索、儲存、維護更加不容易，因此本研究提出不同資料的影響力指標的建立流程。本研究利用行政院的施政方針及施政報告為例，透過本論文提出的流程，經過研究資料的探索與預處理、資料間相似度的探索、資料影響力評估指數、資料視覺化及資料模型設計、及資料查詢模型設計， 使用者可視覺化檢視施政方針及施政報告關聯程度，了解施政報告與施政方針的符合程度，有效控管技術資源應用外，讓管理者更容易追蹤政策綱領與計畫方向是否一致，避免不合乎時宜之政策。

關鍵字

施政方針 ； 施政報告 ； 資料影響力

並列摘要

In recent years, public awareness and participation bring about significant development of open data. People can therefore participate in management and provide decision making ideas; moreover, the transparency of information also speed up the policy push. However, in the era of information explosion, hundreds of thousands of data are produced in every second, which increases the complexity of data querying and database maintaining. The main contribution of this paper is dedicated to enhancing interoperability of policy objective and policy address. The data processing flow proposed in this paper are as follows: data exploration and pretreatment, similarity between data exploration, data impact assessment index, data visualization as well as data model design, and data query model design. Following above flows, we are able to compare the relationship among policy address from different departments by reading their weighted value, which provides reference in data visualization. On the other hand, managers can follow up policy projects more easily, canceling out of date policy projects by checking if they are consistent with policy principles.

並列關鍵字

Policy Objective ； Policy Address ； Data Impact