一种基于维规约的属性加权朴素贝叶斯算法

A attribute weighted naive Bayes algorithm based on the dimension reduction

朴素贝叶斯算法是一种简单而高效的分类算法,但它的属性的条件独立性假设在现实中往往不成立,而且算法本身对高维数据不敏感,如何提高高维数据的分类性能是一个重要的问题。通过确定权重系数进行算法改进,用改进的算法对基于条件信息熵、主成分分析和独立成分分析处理的数据进行分类,并分析性能。

Naive Bayes algorithm is a simple and effective classification algorithm, but its properties of conditional independence assumption in reality tend do not set up, and the algorithm itself is not sensitive to high-dimensional data, how to improve the classification performance of high-dimensional data is an important research problem. Using the improved algorithm processing of data based on conditional information entropy and principal component analysis and independent component analysis classification,and analyzes its performance.