A Study on Normalization Techniques for Privacy Preserving Data Mining
A Study on Normalization Techniques for Privacy Preserving Data Mining
复制标题
DOI：
--
复制
发表时间：
2013
期刊：
影响因子：
--
通讯作者：
C.Saranya;G.Manikandan
中科院分区：
文献类型：
--
作者：
C.Saranya;G.Manikandan
-数据挖掘是一种流行的技术，它从大数据集中提取不熟悉的吸引人的模式。提取的事实被用于各种领域，如市场营销、天气预报和医疗诊断。非常重要的一点是，当组织开始为挖掘过程共享数据时，数据会被曝光，隐私可能会受到侵犯。在许多数据挖掘应用中，隐私正成为一个越来越重要的问题。隐私保护技术为解决这一问题提供了一条新的途径。它在不影响原始数据值的情况下提供合法的数据挖掘结果，从而保证隐私和准确性。在本文中，我们分析了使用规范化技术来实现隐私。我们对这些技术的结果进行了比较，从实验结果可以看出，最小-最大归一化具有最小的误分类误差。
- Data mining is a prevailing technique which extracts the unfamiliar appealing patterns from large data sets. The extracted facts are utilized in various domains like marketing, weather forecasting, and medical diagnosis. It is very vital that the data gets exposed when the organizations start sharing the data for the mining process and privacy may be breached. Privacy is becoming a more and more significant issue in many data mining applications. Privacy preserving techniques gives a new track to solve this problem. It gives legitimate data mining outcomes without edifying the original data values and thus guarantees privacy as well as accuracy. In this paper we have analyzed the use normalization techniques in achieving privacy. We have compared the results of these techniques and from the experimental outcome it can be concluded that Min-Max normalization have minimum misclassification error.