CAJ | 학술논문

针对信息安全检测中的检测精度低的问题，提出了基于稀疏距离入侵特征表达的信息安全检测算法。在该算法中，引入稀疏表达对完备词典进行编码，使得学习的稀疏系数可以具有较好的重构特征；其次利用K-SVD算法和支持向量机进行样本分类训练，使得稀疏特征为样本输入；最后利用粒子群算法对多维测试数据进行粒子映射，在满足适应度函数的条件下进行分类迭代寻优。实验表明，该算法具有较好的检测率。
침대신식안전검측중적검측정도저적문제，제출료기우희소거리입침특정표체적신식안전검측산법。재해산법중，인입희소표체대완비사전진행편마，사득학습적희소계수가이구유교호적중구특정；기차이용K-SVD산법화지지향량궤진행양본분류훈련，사득희소특정위양본수입；최후이용입자군산법대다유측시수거진행입자영사，재만족괄응도함수적조건하진행분류질대심우。실험표명，해산법구유교호적검측솔。
According to the information safety testing in low precision of the question, proposes an information security detection algorithm based on sparse distance invasion features expression. In this algorithm, the introduction of sparse ex-press the complete dictionary for coding, makes the study of sparse coefficient can be better reconstructed characteristic;secondly by K- SVD algorithm and support vector machine (SVM) classification for sample training, making the sparse feature for sample input;finally, using the particle swarm algorithm for multidimensional test data in particle mapping, in meet the fitness function under the condition of classification iterative optimization. Experiments show that this algorithm has good detection rate.