On combining classifiers
On combining classifiers
复制标题
DOI：
10.1109/34.667881
复制
发表时间：
1998-03-01
影响因子：
23.6
通讯作者：
Matas, J
中科院分区：
文献类型：
--
作者：
Kittler, J;Hatef, M;Matas, J
我们开发了一个通用的理论框架，用于组合使用不同模式表示的分类器，并表明许多现有的方案可以被视为复合分类的特殊情况，其中所有模式表示都被联合使用以做出决策。各种分类器组合方案的实验比较表明，在最严格的假设条件下开发的组合规则-求和规则优于其他分类器组合方案。对各种方案对估计误差的敏感性分析表明，这一发现在理论上是合理的。
We develop a common theoretical framework for combining classifiers which use distinct pattern representations and show that many existing schemes can be considered as special cases of compound classification where ail the pattern representations are used jointly to make a decision. An experimental comparison of various classifier combination schemes demonstrates that the combination rule developed under the most restrictive assumptions-the sum rule-outperforms other classifier combinations schemes. A sensitivity analysis of the various schemes to estimation errors is carried out to show that this finding can be justified theoretically.