Implementacija grafičnega uporabniškega vmesnika za izvedbo osnovnih Bayesovih statističnih analiz

Although Bayesian statistics offers a coherent and intuitive framework for statical modeling that allows for explicit uncertainty estimation and seamless integration of prior knowledge, its adoption is often hindered by the required fluency in specialized probabilistic programming languages (such as Stan for example). This thesis presents a graphical user interface (GUI) that allows users to carry out fundamental Bayesian analysis, such as parameter estimation, hypothesis testing, regression and hierarchical modeling, without the need for coding. Implemented as modular widgets in Orange Data Mining and integrated with Stan via the cmdstanpy interface, the system provides an intuitive interface for specifying priors, performing diagnostics and visualization. Testing confirmed stable posterior estimates for several practical examples via a user-friendly interface, making Bayesian methods accessible for education and applied research through computationally robust and transparent tools.