推薦系統應用於遏止走私攻擊以捍衛國家安全;Applying Recommender Systems to Defend National Borders from Smuggling Attacks

NCUIR > School of Management at National Central University > Graduate Institute of Business Administration > Electronic Thesis & Dissertation > Item 987654321/59885

Please use this identifier to cite or link to this item: http://ir.lib.ncu.edu.tw/handle/987654321/59885

Title:	推薦系統應用於遏止走私攻擊以捍衛國家安全;Applying Recommender Systems to Defend National Borders from Smuggling Attacks
Authors:	溫志皓;Wen,Chih-Hao
Contributors:	企業管理學系
Keywords:	國境安全;國家安全;決策樹;貝氏網路;走私;馬可夫覆蓋;Bayesian networks;Decision trees;Homeland security;Markov blanket;National border security;Smuggling
Date:	2013-02-27
Issue Date:	2013-06-19 15:21:36 (UTC+8)
Publisher:	國立中央大學
Abstract:	在台灣四面環海的地理條件下，海上交通及運輸甚為發達。同樣地，經由海上走私的項目與金額，也非常龐大。因為走私而進入國內的毒品、槍械、非法移民或情治人員將會嚴重影響國內治安及國家安全。海巡署所屬各單位在港口對進出船隻執法檢查是遏止走私事件的重要攔截防線。目前現行作業方式是在船隻出港或進港的同時，依岸巡人員之經驗來選擇抽檢的船隻。這種方式往往會依人員更迭與累積的經驗多寡而影響查緝走私效率的良窳。此外，走私項目的差異(如漁獲與槍械)，對於治安將會造成更懸殊的影響。本研究為能提高對走私船隻的鑑別率，將針對走私項目(如漁獲、槍械、毒品…等等)與蒐集變數(如母港、船主年齡、進出港口…等等)之間其相依或是獨立的關係，採用分類樹與貝氏網路演算法，分別建立變數相依分類樹、變數獨立分類樹、以及變數相依貝氏網路、變數獨立貝氏網路等四種模型。尤其，本研究考慮走私項目將會對社會治安產生的危害價值納入分析內容，提供建議檢查的船隻選擇方式。因此，本研究的結果可在相同的檢查時間下，對走私船隻有更準確的鑑別率，同時，選取走私價值最大的船隻以降低對社會治安與國家安全可能造成的損害。Being an island state, Taiwan’s maritime traffic and sea transportation is well-developed; however, smuggling by sea converts into a critical issue. Coast Guard Administration (CGA) becomes a vital intercept line to curb smuggling events for Taiwan’s social and national security. Yet, it is dubious and untrustworthy for current practices since inspectors of CGA check vessels selectively by personnel experience. In addition, the repercussion caused by different smuggling items (such as fishery and guns) is varied.In order to improve the identification rate of smuggling vessels for smuggling items (such as fishery, guns, drugs etc.) and collecting variables (such as the home port, the age of boat owners, the vessels’ entry and departure etc.) between relationships of dependence or independence, this research applies classification trees and Bayesian network algorithms respectively creating four models: a dependent classification tree model, an independent classification tree model, a dependent Bayesian network model, and an independent Bayesian network model. In particular, this research offers an inspective method for vessel’s selection owing to the impact of social security caused by smuggled items. Therefore, this research on one hand improves the identification accuracy of smuggling vessels, on the other, reduces the possible damage in social and national security by selecting the vessels with the largest smuggling value.
Appears in Collections:	[Graduate Institute of Business Administration] Electronic Thesis & Dissertation

Files in This Item:

File	Description	Size	Format
index.html		0Kb	HTML	1119	View/Open

社群 sharing

Loading...