This thesis addresses the challenge to provide personalized offers to users. The level of personalization directly depends on the quality of the matchmaking process, i.e. finding those offers that are most attractive to a particular user based on his or her interests. We study the matchmaking process in the context of e-tourism. The goal is to support tourists in their decision-making during the pre-trip phase and to facilitate the process of identifying those tourism objects of a specific destination that best fit the tourists' preferences. To achieve this goal, we propose an iterative matchmaking process that matches tourist profiles with the characteristics of tourism objects in order to obtain a ranked list of appropriate tourism objects for a particular tourist. The matchmaking process is composed by two main steps. For the first step, we devise a stereotype approach based on tourist types to model a basic user profile, reflecting tourists¿ preferences according to travel-related categories (e.g., culture, sightseeing, sports or nature). This profile is then related against the generic characteristics of the tourism objects in order to recommend a top-N list of tourism objects. The generic characteristics of the tourism objects are modelled based on the same typology. To continuously enhance the quality of the matchmaking process, we exploit tourist feedback to dynamically adapt and refine tourist profiles (e.g., a tourist may be a cultural type but may express a dislike of museums). This is achieved by the second step. Its task is to consider positive/negative tourist feedback in order to derive specific interests and to re-adapt the matches between one particular tourist and the set of relevant tourism objects by taking into account his/her specific interests. We develop a tourism ontology and use it as basis in order to model the specific interests of the tourists and the specific attributes of the tourism objects. The first and second matchmaking steps can be combined. Tourists can criticize the proposed items by stating positive/negative feedback, which will be used to refine their profile and to deliver a new set of tourism objects. As long as they are not satisfied with the recommendations, they can repeat this process. Our approach is tested through a prototypical recommendation system that recommends tourists visiting Vienna appropriate tourism attractions that are tailored to their personal needs. We conduct a user study by asking users to interact with the system and fill in a questionnaire afterwards.