引入高阶累积量的数据库访问特征选择算法

在进行数据库访问的过程中,由于受到很多不确定性因素的干扰,使数据库中存在大量噪声,影响了数据库访问的效率。提出一种引入高阶累积量的数据库访问特征选择算法,依据高阶累积量两个统计独立随机过程之和的累积量等于各个随机过程累积量之和的性质,对数据库进行去噪处理。在此基础上,采用SVM无监督算法实现数据库访问特征选择。仿真实验结果表明,采用所提算法进行数据库访问特征选择,不仅具有较高的特征选择精度,而且特征选择效率也明显高于传统算法,同时特征选择结果所含冗余特征低于传统算法,验证了所提算法在数据库访问特征选择方面的性能。

Introduction of high-order cumulant database access feature selection algorithm

In the process of the database access, due to the interference of many uncertain factors, there are a lot of noise in the database, which would influence the efficiency of database access, and put forward a kind of introduction of high-order cumulant database access feature selection algorithm, based on higher order cumulants of the sum of two statistically independent random process cumulant is equal to the sum of the stochastic process cumulant, the nature of the database to deal with the noise, on this basis, using the SVM unsupervised algorithm database access feature selection. The simulation results show that the proposed algorithm for database access feature selection, not only has higher precision of feature selection, and also significantly higher than that of traditional algorithm, feature selection efficiency and feature selection results contain redundancy feature is lower than the traditional algorithm, the proposed algorithm is verified in the database access feature selection in terms of performance.#$NLKeywords: higher order cumulant;Database;Feature selection