以知識本體為基礎建構病毒分類知識系統

An Ontology-based Approach towand Knowledge Systems in Classification of Virus

在網際網路時代裡，形形色色的網路應用在無地域限制下快速的發展。同時，基於某種原因，電腦病毒開發者，也將不斷開發出新型的變種電腦病毒，藉由網際網路的便利性四處散佈。當使用者遭受電腦病毒感染與威脅時，快速識別出電腦病毒的型態，並找尋相關解決方案是非常重要的。本研究提出藉由知識本體結合機器學習的創新方式，進行電腦病毒的分類，具體而言，本研究採用自我組織映射網路（self-organizing map）結合 k平均法（k-means）的自動分群技術，並導入 C4.5決策樹演算法以建立電腦病毒的知識階層架構。此種方式能夠以客觀的方式迅速有效的自動化塑模電腦病毒知識本體，並能透過本體規則推論找出最適合的病毒感染解決方案。

In the era of Internet, many varied information flows prompt a rich of usage over Internet without geographical limitation. However, for some purpose, many computer virus creators also take the benefit of Internet and make varied viruses diffuse ubiquitously on the Internet. It should be necessary to recognize the virus type and look for solutions immediately when a computer is attacked by a virus. Based on the integration of ontology learning and machine learning, this research proposes a useful solution to classify viruses to their associated categories. More specifically, we propose a novel method, which integrates self-organizing map (SOM) with K-means into C4.5 decision tree in order to objectively produce a hierarchical knowledge base for virus. Our proposed method is able to build the computer virus ontology automatically, effectively and efficiently and find the most suitable solution for an infected computer via rule inference in this computer virus ontology.