基于决策树的数据库脆弱水印算法研究

Fragile Watermarking Algorithm Based on Decision Tree

为了检测数据库篡改以保护数据库版权,对传统的数据库数字水印技术进行分析,提出一种从引入三角模糊函数优化阈值划分和建立子树剪割原则进行改进的C4.5决策树算法,并采用改进型C4.5决策树进行数据库元组分组并进行hash函数变换通过主键排序,从而实现水印嵌入和元组数据的检测和篡改定位,同时在各属性中嵌入水印和校验码,进而实现水印嵌入并在检测中完成篡改属性定位,最后通过定性分析中发现算法具有更广适用性和有效进行篡改检测和定位。

In order to detect tampering with the database copyright protection to protect the database, analysis of database digital wa termarking technology of the tradition, proposes a from introducing triangle fuzzy function optimization threshold division and eslablishment of tree cutting principle of improved C4. 5 decision tree algorithm, and the database tuples packet and hash transform by primary key order by using the improved （24. 5 decision tree, so as to realize the watermark embedding and tuple data detection and localization, and in the proper ties of the embedded watermark and the check code, and then realize the watermark embedding and tamper with the attribute location in de tection, finally through the qualitative analysis discovered that the algorithm has a wider applicability and effectiveness of tamper detection and localization.