雲端環境上的複合式知識融合與推論

Hybrid Knowledge Fusion and Inference on Cloud Environment

張景棠(CHING-TANG CHANG)

指導教授 ： 張玉山 ； 戴志華

國立臺北大學/電機資訊學院/資訊工程學系/碩士(2016年)

摘要

在大數據的時代，數量驚人的知識無時無刻並且以各式各樣的方法在產生。知識的融合與推論成為了應用知識的重要議題。
在這篇論文中，我們關注於可以表達為關聯式規則的複合式知識（包含類別或是數值的資訊），並且討論在雲端環境上的複合式知識融合與推論的相關問題。我們一共列舉了6個議題並且提出一個有效的解決辦法：HyKFICE(Hybrid Knowledge Fusion and Inference on Cloud Environment)。HyKFICE可以利用機率理論來對已知發生的實情進行知識的融合與推論，來推測未來可能發生的事件及其機率。HyKFICE利用將相似的知識轉換成 3-layer directed bipartite graph來達到雲端上的平行運算。在實驗部份利用真實的資料來展示效能，以及系統的執行時間不只受資料量大小的影響，也會因為資料之間的關聯性而有變化。

關鍵字

知識融合 ； 知識推論 ； 機率模型

並列摘要

In the age of big data, an incredible amount of knowledge is produced everywhere everyday through various ways.
Knowledge fusion and inference thus has become an important issue for better utilization of knowledge.
In this paper, we focus on the hybrid knowledge that can be represented in the form of association rules (with categorical and/or numerical information), and address the problem of such hybrid knowledge fusion and inference on cloud environment.
In light of the use of the problem, we specify six issues of knowledge fusion and inference and propose a HyKFICE (Hybrid Knowledge Fusion and Inference on Cloud Environment) system as an effective solution. HyKFICE is capable of inferring the possibilities of the happening of events at a given condition through knowledge fusion and inference based on the probability theory.
HyKFICE can also perform the computation in parallel on clouds by grouping and summing up similar knowledge in 3-layer bipartite graphs.
Experiments conducted on real data sets demonstrate the efficiency of HyKFICE and show that
it is not only the amount of knowledge but also the associations between knowledge dominating the execution time of the hybrid knowledge fusion and inference.

並列關鍵字

Knowledge fusion ； Knowledge inference ； Probabilistic model

參考文獻

adults with acute myocardial infarction. J. Medical Systems, 2013.

[12] Xin Luna Dong, Evgeniy Gabrilovich, Geremy Heitz, Wilko Horn, Kevin Murphy, Shao-

[22] Xin Dong, Evgeniy Gabrilovich, Geremy Heitz, Wilko Horn, Ni Lao, Kevin Murphy,

[45] Eibe Frank, Mark A. Hall, Geoffrey Holmes, Richard Kirkby, and Bernhard Pfahringer.

[46] Mark A. Hall, Eibe Frank, Geoffrey Holmes, Bernhard Pfahringer, Peter Reutemann, and