Recommender system for B2B data

This bachelors thesis examines the topic of practical recommender systems. The document consists of both theoretical and practical parts. The theoretical part starts by describing and defining the concept of a recommender system. Then the history and development of these systems are briefly described. The main concepts and design approaches are introduced based on several publications about the topic. The strengths and weaknesses of each design approach are considered. One of these approaches is chosen, specifically the content-based approach, with the help of the facts described in the theoretical part. The principles of this approach are used to design a practical recommender system that proposes potential customers to sales representatives in the B2B market. The input of this algorithm are real-world data from the Czech market provided by a company with experience and expertise in this field. The first result of the practical part of the thesis is an automated algorithm that recommends companies to sales representatives based on their historical behavior. The second result is a web interface created to configure and test the algorithm. The last part of the thesis evaluates the created solution and proposes future steps to further develop and improve it.