自民國成立後,國家重要慶典或哀悼等活動曾實施多次減刑,每一次的減刑均受到國人高度的關注。從社會觀點來看,一次釋放大批受刑人可能會增加社會的不安定因素,因此減刑制度有必要透過相關研究進一步分析、評估,使決策者在思考減刑的實施能更周全。減刑出獄的受刑人,若能改過向善,政府實施之減刑方能有其正面成效;但若減刑出獄人再犯率甚高,減刑成效勢必大打折扣。因此,減刑犯再犯因素的探討極為重要。 過往的研究較常採用統計方法來分析,本研究擬以資料探勘技術探討減刑議題,透過約略集合理論篩選出核心屬性子集合,並結合關聯法則原理,找出影響再犯規則,挖掘各項罪名顯著之特徵屬性。本研究期望能減少或預防再犯的發生,挖掘各項罪名哪些屬性是關鍵的,提供我國減刑制度決策之參考,如減刑條件、教育訓練或賞罰制度等,藉以降低犯罪率及社會風險。
According to history, celebration and mourning activities have implemented several commutations since the Republic of China was established. Each commutation is subject to a high degree of concern. From the social standpoint, the release of a large number of inmates of the prison might cause social panic. Therefore, further analysis and evaluation of commutation is necessary for decision-makers on the commuted issues.If early release prisons can turn over a new leaf, commutation will have a positive effect. However, high recidivism rate would discourage commutation. Most papers applied statistical methods to analyze the crime factor before. In this project, we utilize data mining methods to investigate aspects of the commutation issues. We integrate rough set theory with association rules to find the major factors for Recidivism. The result of this research is expected to be helpful to the decision of commutation.