محمد حمزه لو

In today's world, football is no longer just a sport, but a meta-sports phenomenon. Football is intertwined with economy, culture, politics and people's lives. According to the statistics of Bein Sports, the 2022 World Cup in Qatar had 5.4 billion visitors. This statistic shows people's interest in this sport. Also, one of the attractions of this sport is the predictions and bets in this field, which has caused the financial turnover of this field to be very high. In this field, data science and data mining are used to increase the accuracy of prediction and thus increase the profit. In this research, two prediction models have been investigated and eva‎luated based on previous results and players' scores with different data mining techniques, and finally, a hybrid model of these two models has been created and presented. In the results of the eva‎luations, it was found that all three models are still weak in predicting draw results, and in the best case, 33% of draw results were correctly recognized in the combined model and support vector machine technique. However, in relation to the prediction of the results associated with the win of the host or guest teams, an approximate accuracy of 80% was achieved with logical regression and random forest techniques.

پيش‌بيني مسابقات فوتبال , داده‌كاوي , بيز ساده لوح , ماشين بردار پشتيبان , جنگل تصادفي , رگرسيون منطقي , نزديك‌ترين همسايه , تقويت گراديان شديد

Soccer Prediction , Data Mining , Naive Bayes , Support Vector Machine , Random Forest , Logistic Regression , Nearest Neighbor , Extreme Gradient Boosting

Mohammad Hamzelu

Dr. Mohammad Fathian