情报杂志Issue(10)：148-152,5.DOI:10.3969/j.issn.1002-1965.2014.10.025
基于随机森林的分类器在犯罪预测中的应用研究
Application of an Improved Random Forest Based Classifier in Crime Prediction Domain
孙菲菲 1曹卓 2肖晓雷3
作者信息
- 1. 中国人民公安大学反恐怖学院 北京 100038
- 2. 中国人民公安大学犯罪学学院 北京 100038
- 3. 中国人民公安大学侦查学院 北京 100038
- 折叠
摘要
Abstract
Crime prediction has always been an outstanding issue for public security department. Based on a branch of Module Combina-tion Classifier--Random Forest classifier, this article introduces a new way of classification which can be applied to crime prediction based on the current situation of the application of Machine Learning technology. We propose a novel Random Forest classifier which is com-posed of a group of degradation decision trees, and decision trees that are more distinctive are selected in the next round classification. Fi-nally the article gives a successful new way of building classifier in crime predication domain, to reach an accurate and effective conclu-sion, and the high reliability of classifier results compared to general random forest classifiers is also demonstrated based on simulation ex-periment results.关键词
随机森林/机器学习/分类器/犯罪预测/决策树/数据挖掘Key words
random forest/machine learning/classifier/crime prediction/decision tree/data mining分类
社会科学引用本文复制引用
孙菲菲,曹卓,肖晓雷..基于随机森林的分类器在犯罪预测中的应用研究[J].情报杂志,2014,(10):148-152,5.
