In this thesis, I organize two independent projects into five chapters. The first chapter introduces Liquid Association and our proposed method to accelerate its computation. The second chapter is related to the design of the computational structure for Liquid Association website (LAP3). The third chapter is regarding the application of Liquid Association to Global Health Observatory (GHO) data. The fourth chapter describes a novel method to model the distribution of human ratings on word-similarity. The last chapter focuses on the analysis of the relationship between the knowledge-based approach and the corpus-based approach.