From Screen to Success: Investigating Machine Learning Algorithms to Predict Movies’ Success, Based on Actor Features

This dissertation uses the large movie datasets available on IMDb to study the predictive potential of Machine Learning algorithms on the critical success of movies, focusing on the role of actor attributes. This study dives into the various factors that contribute to a ''film's success, including actor and movie features as well as movie ratings, given the complex relationship between cinematic success, which is associated with commercial and critical acclaim. The research adopts the CRISP-DM methodology and applies several Machine Learning techniques, with the Random Forest algorithm surging as the most effective in predicting movie ratings based on actor attributes, particularly in the Sci-Fi genre. This dissertation promotes analytics techniques in the motion picture business and provides a resource for industry stakeholders to explore the uncertain world of film production and audience preferences.