【目的/意義】近年來,世界各地的暴恐活動呈現出破壞性大、影響范圍廣、計劃性強、手段多樣等特點,這也決定了打擊恐怖主義要把反恐預警工作作為重中之重。利用數據挖掘技術進行涉恐情報分析,發現涉恐線索是實現反恐預警的有效手段之一。數據預處理是使用各種數據挖掘方法的前提和基礎。【方法/過程】本文將系統研究數據預處理在涉恐情報量化分析中的應用。主要討論數據清洗、數據集成、數據離散化以及數據二元化等預處理方法在涉恐信息處理時應該重點注意的問題。文中所有涉恐數據特征均從媒體公開報道中獲取。【結果/結論】本文討論的預處理方法可以為展開涉恐情報數據挖掘消除數據缺陷,提供標準化的基礎數據,為挖掘結果分析和可視化展現打下良好基礎,有利于反恐預警的實現。

【Purpose/significance】In recent years, violent terrorist activities all around the world display the characteristics of devastating, larger influence, strong planning and various measures. Terror threat warning becomes the most important work to fight against terrorism. Data mining technology plays an important role to accomplish terror threat warning. Data preprocessing is the basis of all data mining technologies.【Method/process】This paper proposes how to make data preprocessing in the field of counter terrorism intelligence quantitative analysis. Four types of preprocessing methods, including data cleaning, data integration, data discretization and data dualization, are discussed to deal with terror related information. All fundamental data are based on public reports.【Result/conclusion】These preprocessing methods could eliminate data defects and provide standardized data for next data mining step to show mining conclusions and visible results.