The European Union’s economy, during the last years, has been influenced a lot due to the development of artificial intelligence (AI). This impact is been growing very fast thanks to the continuous introduction of new technologies and new investments. However, artificial intelligence could be dangerous if not regulated, as it acquires and uses huge amounts of data and could overtake human’s capabilities and skills. As matter of facts, European community still takes a dimmer view about AI, since it does not exist yet a regulatory framework able to completely protect the European citizen. 2020 was a year in which the spread of COVID-19 has affected the entire world, causing a global pandemic. Due to this phenomenon, countries had the need to start monitoring the trend of the contagion throughout the application of a data driven approach. In this way, governments could take action by introducing containment measures. In this research, an econometric model has been applied in order to monitor and predict the development of the contagion over the new daily infected. The model, a Poisson Autoregression, has been applied on a sample composed by those EU countries that have registered at least 200.000 contagions. Thanks to its ability to learn from data, this model is able to detect the trend of contagion’s peaks over the time. With this form, it is possible to overcome the limits of the previous epidemiological approaches used to compute the contagion reproduction’s rate. In particular, to this model are applied coefficients able to analyse the dependence on the number of contagions expected on the short and long term. The results show how the countries in the sample have had similar trends both during the first and the second wave of contagions. Finally, from the calculation of the coefficient of determination for each country, it has turned out that the majority of the nations have registered a value higher than 0.9, indicating good predictive abilities of the Poisson autoregressive model.
L’impatto dell’intelligenza artificiale sull’economia dell’Unione Europea sta assumendo dimensioni sempre più grandi, grazie alla costante introduzione di nuove tecnologie e agli investimenti effettuati dalle varie potenze mondiali. Tuttavia, l’intelligenza artificiale, sfruttando l’utilizzo di grandi quantità di dati e talvolta superando addirittura le capacità umane, ha delle implicazioni di cui bisogna tenere conto e che oggigiorno sono viste con scetticismo da una buona parte della comunità Europea, anche perché non esiste ancora un quadro normativo capace di tutelare completamente il cittadino. Il 2020 è stato un anno segnato dalla diffusione del virus COVID-19 e in tutto il mondo è sorta la necessità di monitorare l’andamento del contagio tramite un data driven approach per permettere ai governi di prendere decisioni in merito alle misure di contenimento del virus ed evitarne la diffusione. A tal proposito questa tesi applica un modello econometrico capace di monitorare e prevedere la dinamica dell’evoluzione del contagio sulla base dei nuovi infetti giornalieri. Il modello, che è una auto regressione di Poisson, è stato applicato su un campione costituito dai paesi dell’Unione Europa che hanno registrato un numero di contagi maggiore di 200.000. Grazie alla sua abilità di imparare dai dati, il modello è capace di rilevare l’andamento e i picchi del contagio nel tempo. Inoltre, grazie all’utilizzo di coefficienti capaci di analizzare la dipendenza sul numero di contagi previsto nel breve e nel lungo termine, è possibile superare i limiti dei modelli epidemiologici usati in precedenza nel calcolo dell’indice di riproduzione del contagio. I risultati ottenuti mostrano come i paesi sotto osservazione abbiano seguito andamenti che si rispecchiano nella prima e seconda ondata del contagio. Infine, calcolando il coefficiente di determinazione per ogni paese, è emerso che la maggioranza delle nazioni a cui è stato applicato il modello ha registrato valori sopra 0.9, dimostrando le sue buone capacità predittive.
L'impatto dell'intelligenza artificiale sull' UE: un alleato contro la SARS-CoV-2
ALACCHI, DAVIDE
2019/2020
Abstract
The European Union’s economy, during the last years, has been influenced a lot due to the development of artificial intelligence (AI). This impact is been growing very fast thanks to the continuous introduction of new technologies and new investments. However, artificial intelligence could be dangerous if not regulated, as it acquires and uses huge amounts of data and could overtake human’s capabilities and skills. As matter of facts, European community still takes a dimmer view about AI, since it does not exist yet a regulatory framework able to completely protect the European citizen. 2020 was a year in which the spread of COVID-19 has affected the entire world, causing a global pandemic. Due to this phenomenon, countries had the need to start monitoring the trend of the contagion throughout the application of a data driven approach. In this way, governments could take action by introducing containment measures. In this research, an econometric model has been applied in order to monitor and predict the development of the contagion over the new daily infected. The model, a Poisson Autoregression, has been applied on a sample composed by those EU countries that have registered at least 200.000 contagions. Thanks to its ability to learn from data, this model is able to detect the trend of contagion’s peaks over the time. With this form, it is possible to overcome the limits of the previous epidemiological approaches used to compute the contagion reproduction’s rate. In particular, to this model are applied coefficients able to analyse the dependence on the number of contagions expected on the short and long term. The results show how the countries in the sample have had similar trends both during the first and the second wave of contagions. Finally, from the calculation of the coefficient of determination for each country, it has turned out that the majority of the nations have registered a value higher than 0.9, indicating good predictive abilities of the Poisson autoregressive model.È consentito all'utente scaricare e condividere i documenti disponibili a testo pieno in UNITESI UNIPV nel rispetto della licenza Creative Commons del tipo CC BY NC ND.
Per maggiori informazioni e per verifiche sull'eventuale disponibilità del file scrivere a: unitesi@unipv.it.
https://hdl.handle.net/20.500.14239/561