k-Anonymity and differential privacy can be considered examples of Boolean definitions of disclosure risk. In contrast, record linkage and uniqueness are examples of quantitative measures of risk. Record linkage is a powerful approach because it can model different types of scenarios in which an adversary attacks a protected database with some information and background knowledge. Transparency holds in data privacy when data is published together with details on their processing. This includes the data protection method used and its parameters. Intruders can use this information to improve their attacks. Specific record linkage algorithms can be defined to take into account this information, and to define more accurate disclosure risk measures. Machine learning and optimization techniques also permits us to increase the effectiveness of record linkage algorithms. This talk will be focused on disclosure risk measures based on record linkage. We will describe how we can improve the performance of the algorithms under the transparency principle, as well as using machine learning and optimization techniques.