摘要
针对模式识别中的多分类器集成,通过挖掘测试样本特征属性的相关性,结合训练集的条件独立性分析对每个样本赋予分类规则,构造分类森林(而非单个决策树)进行模型集成。整个学习过程能够自适应确定各决策树结构和数量,并充分发挥集成模型的伸缩性和扩展性。在UCI机器学习数据集上的实验结果验证了本方法的有效性。
For the multiple classifier integration in a decision tree was built to realize the submodel attributes in the test sample and conditional independence analysis trees can be defined adaptively d the pattern recognition, integration by mining th a decision forest rather than e relevance in the predictive giving the distinct classification rule to each sample based on the of the training set. The structure and the number of the decision uring the learning process. Experiments on UCI learning data sets proved the feasibility and effectiveness of the proposed method.
出处
《吉林大学学报(工学版)》
EI
CAS
CSCD
北大核心
2010年第1期155-158,共4页
Journal of Jilin University:Engineering and Technology Edition
基金
国家自然科学基金项目(60275026
60803055)
关键词
人工智能
模式识别
决策森林
条件独立性假设
数据挖掘模型
artificial intelligence
pattern recognition
decision forest
conditional independence assumption
data mining model