| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 4.4 MB | Adobe PDF |
Authors
Advisor(s)
Abstract(s)
Atualmente, os dados simbolizam a matĂ©ria-prima mais valiosa do planeta. AtravĂ©s deles, Ă© possĂvel aceder a informaçÔes vitais para a sobrevivĂȘncia e sucesso das empresas de todo mundo. Neste contexto, a InteligĂȘncia Artificial (IA) assume um papel preponderante na obtenção de informação valiosa âescondidaâ nas diversas formas que os dados podem assumir. Feitas estas consideraçÔes, e sabendo das proporçÔes que a importĂąncia da IA estĂĄ a tomar nos Ășltimos anos no processo de tomada de decisĂŁo, foi desenvolvido este projeto, no Ăąmbito da unidade curricular Projeto/Dissertação/EstĂĄgio (PROJIA). Este visa a anĂĄlise dos dados existentes no mundo das apostas desportivas, para que, atravĂ©s destes, seja possĂvel retirar informação valiosa para o aumento do sucesso dos prognĂłsticos em jogos de futebol. Para este efeito, recorreu-se a dados extraĂdos atravĂ©s de uma Application Programming Interface (API), os quais foram, posteriormente, tratados para servirem de base de informação do modelo. determinando este, por sua vez, o vencedor num determinado jogo de futebol, atravĂ©s de algoritmos de Deep Learning (DL). Posto isto, com este modelo, foi possĂvel constatar que a probabilidade de acertar o vencedor de um jogo de futebol aumenta em ligas com um maior nĂșmero de jogos que ocorram de forma mais regular. JĂĄ em campeonatos em que participem clubes de diferentes paĂses, e em que os jogos entre as equipas encerrem um grande intervalo temporal, o modelo nĂŁo Ă© capaz de acertar tĂŁo eficazmente na equipa vencedora.
Nowadays, data is one of the most valuable raw materials on the planet. Through them is possible to obtain information vital for the viability and success of companies around the world. In this context, Artificial Intelligence (AI) assumes a leading role in obtaining valuable information "hidden" in the various forms that data can have. Having made these considerations and knowing the proportions that AI is portraying in the decision-making process, the importance of this project, that was developed in the scope of the curricular unit Project/Dissertation/Internship (PROJIA), can be appraised. Aiming at the analysis of existing data in the world of sports betting, so that, through these, it is possible to extract valuable information to increase the success of predictions in football games. For this purpose, data extracted through an Application Programming Interface (API) was used and then processed to feed the model's database and consequently is used to determine the winner in football matches through Deep Learning (DL) algorithms. With this modelâs output, it was possible to observe that the probability of determining the winner of a football match is superior in leagues with a larger number of matches that occur more regularly. On the other hand, in leagues where clubs from different countries participate, and where the matches span over a larger time interval, the model is not able to determine the winning team as effectively.
Nowadays, data is one of the most valuable raw materials on the planet. Through them is possible to obtain information vital for the viability and success of companies around the world. In this context, Artificial Intelligence (AI) assumes a leading role in obtaining valuable information "hidden" in the various forms that data can have. Having made these considerations and knowing the proportions that AI is portraying in the decision-making process, the importance of this project, that was developed in the scope of the curricular unit Project/Dissertation/Internship (PROJIA), can be appraised. Aiming at the analysis of existing data in the world of sports betting, so that, through these, it is possible to extract valuable information to increase the success of predictions in football games. For this purpose, data extracted through an Application Programming Interface (API) was used and then processed to feed the model's database and consequently is used to determine the winner in football matches through Deep Learning (DL) algorithms. With this modelâs output, it was possible to observe that the probability of determining the winner of a football match is superior in leagues with a larger number of matches that occur more regularly. On the other hand, in leagues where clubs from different countries participate, and where the matches span over a larger time interval, the model is not able to determine the winning team as effectively.
Description
Keywords
InteligĂȘncia Artificial Apostas Desportivas Machine Learning Deep Learning Random Forest Classifier LSTM CNN Artificial Intelligence Sports Betting