**4. Discussion**

The fact that the digitalization process has become mandatory all over the world with Covid-19 pandemic has accelerated the data flow. It has become even more important to collect the necessary data, analyze it correctly and reveal reliable information. This situation has triggered the use of data mining methods to increase productivity and provide high quality products/services in almost all sectors. While applying data mining methods, it is obvious that if privacy is not taken into consideration during the data life cycle, irreversible damages will occur for individuals/institutions and organizations.

In order to increase the access and benefits of data mining technology, before applying PPDM techniques, "privacy" should be defined precisely, measurement metrics should be determined and the results obtained should be evaluated with these metrics. For this reason, this study primarily focused on the definition of privacy. The term privacy is quite extensive and does not have a standard definition. It is quite challenging in measuring privacy, as there is no standard privacy definition. Some measurement metrics are mentioned in this chapter, but metrics are usually determined by application. The lack of a standard privacy measurement metric also make challenging the comparison and evaluation of the developed PPDM techniques.

In the age of digital and online business, privacy protection needs to be done at the individual and organizational levels. Privacy protection at the individual level depends on person who is influenced by religious beliefs, community norms and culture. For this reason, the concept of personalized privacy, which allows individuals to have a certain level of control over their data, has been proposed. However, it has been observed that there are difficulties in implementing personalized privacy, as people think that compromising their privacy for applications they think is well-intentioned will not damage. Therefore, in the context of personalized privacy, new solutions are required for the trade-off between privacy and utility.

To effectively protect organizational level data privacy [7]; Policy makers in organizations should support privacy-enhancing technical architectures/models to securely collect, analyze and share data. Laws, regulations and fundamental principles regarding privacy should be analyzed by organizations. It is necessary for organizations to include the data owners in their assessment of privacy and security practices. Data owners should involve the whole process about what data is collected, how it is analyzed and for what purpose it is used. In addition, they should have the right to correct personal data in order to avoid negative consequences of incorrect data. Organizations should employ data privacy analysts, data security scientists, and data privacy architects who can develop data mining applications securely.

From a technical point of view, methods that protect confidentiality in data analytics are still in their infancy. Although studies continue by different scientific communities such as cryptography, database management and data mining, an interdisciplinary study should be conducted on PPDM. For example, the difficulties encountered in this process should also be addressed from a legal perspective. Thus, a better roadmap for next-generation privacy-preserving data mining design can be developed by academic researchers and industrial practitioners.
