摘要
贝叶斯决策理论是统计模式识别中的一个基本方法。依据贝叶斯决策理论设计的分类器具有最优的性能,即所实现的分类错误率或风险在所有可能的分类器中是最小的,因此经常被用来衡量其他分类器设计方法的优劣。贝叶斯决策是一个很有效的分类工具,但它仍然存在着一定的错误率和风险,因此还需进一步的改善和完善。
Bayesian decision theory is a basic method of Statistical Pattern Recognition. The classification tools designed according to bayesian decision theory have optimal performance, i. e. , classification error rate or risks of the result are the lowest among all possible classification tools, so are frequently used to be the touchstones of other design method. Bayesian decision theory is a very effective classification tool,but it still bring in error rate and risks. So bayesian decision theory should be further improved.
出处
《北京工业职业技术学院学报》
2008年第2期7-10,共4页
Journal of Beijing Polytechnic College
关键词
贝叶斯决策理伦
最优
分类错误率
分类工具
Bayesian decision theory
optimal
classification error rate
classification tools
