Articolo su International Journal of Data Science and Analytics

È stato publicato in open access sulla rivista 𝗜𝗻𝘁𝗲𝗿𝗻𝗮𝘁𝗶𝗼𝗻𝗮𝗹 𝗝𝗼𝘂𝗿𝗻𝗮𝗹 𝗼𝗳 𝗗𝗮𝘁𝗮 𝗦𝗰𝗶𝗲𝗻𝗰𝗲 𝗮𝗻𝗱 𝗔𝗻𝗮𝗹𝘆𝘁𝗶𝗰𝘀 (Springer) l’articolo “𝘖𝘱𝘵𝘪𝘮𝘢𝘭 𝘱𝘰𝘭𝘪𝘤𝘺 𝘭𝘦𝘢𝘳𝘯𝘪𝘯𝘨 𝘸𝘪𝘵𝘩 𝘰𝘣𝘴𝘦𝘳𝘷𝘢𝘵𝘪𝘰𝘯𝘢𝘭 𝘥𝘢𝘵𝘢 𝘪𝘯 𝘮𝘶𝘭𝘵𝘪-𝘢𝘤𝘵𝘪𝘰𝘯 𝘴𝘤𝘦𝘯𝘢𝘳𝘪𝘰𝘴: 𝘦𝘴𝘵𝘪𝘮𝘢𝘵𝘪𝘰𝘯, 𝘳𝘪𝘴𝘬 𝘱𝘳𝘦𝘧𝘦𝘳𝘦𝘯𝘤𝘦, 𝘢𝘯𝘥 𝘱𝘰𝘵𝘦𝘯𝘵𝘪𝘢𝘭 𝘧𝘢𝘪𝘭𝘶𝘳𝘦𝘴” curato da Giovanni Cerulli.

The paper discusses: Offline and online Optimal Policy Learning (OPL) | Multi-action decision settings | Risk-adjusted policy learning | Regret minimization | The role of overlap and unconfoundedness | Applications to data-driven public policy design