示例学习算法IBLE和ID3的比较研究
Comparative Study of IBLE and ID3 Algorithms for Learning From Examples
-
摘要: 为了比较研究IBLE算法和ID3算法的学习性能,本文用大量的质谱数据对两种算法做了学习实验。经过学习,IBLE 的平均预测率为93.96%。而ID3为81.76%,而且IBLE 获得的知识在表示和内容上与专家知识具有较高一致性。文中对两种算法出现上述差异的原因进行了理论分析。Abstract: The learning performance of IBLE and ID3 algorithms is compared and studied with a great number of mass spectra.The average predictive accuracy is 93.96% for IBLE and 81.76% for ID3.The knowledge obtained by IBLE is quite identical with expert knowledge in the representation and contents.The reason for the above differences between IBLE and ID3 is discussed in theory.
下载: