Views
41
Downloands
12
Bu çalışma yapay sinir ağları (YSA) modeli kullanılarak Türkiye Süper Lig sezon sonu takım sıralamasının, atılan ve yenilen gol sayısı giriş değişkenlerine göre tahmin edilmesi amacıyla yapılmıştır. Çalışma kapsamında Türkiye Süper Liginde 2015/2016, 2016/2017 ve 2017/2018 sezonlarında oynanan toplam 918 maçta atılan ve yenilen gol sayısı değişkenlerine ait veriler değerlendirilmiştir. Türkiye Süper Liginde 2015/2016 ve 2016/2017 sezonlarında oynanan maçların analizi yapılarak 2017/2018 sezon sonu lig sıralaması tahmin edilmiştir. Çalışmada değerlendirilen veriler eğitim ve test için rastgele yöntemle ayrılmıştır. Takımların lig sıralaması 0 (sıfır) ile 1 (bir) aralığındaki sayısal değerlerle modellenmiştir. Geliştirilen YSA modeli ile yapılan analizlere göre Türkiye Süper Lig takım sıralaması birçok takım için (test veri kümesi) % 99’un üzerinde doğruluk oranıyla tahmin edilmiştir. Türkiye Süper Liginde sezon sonu takım sıralamasını atılan ve yenilen gol sayılarının doğrudan etkilediği belirlenmiştir. Futbolda sezon sonu takım sıralamasının makine öğrenme yöntemi ile tahmin edilmesi, kulüplerin sezon sonu lig sıralamasında hedefledikleri yerlere göre transfer politikaları belirlemelerini sağlayabilir.
This study was done using the artificial nerve networks (YSA) model in order to estimate the team ranking of the final season of the Turkish Super League, the number of goals thrown and defeated according to the input variables. In the framework of the study, the total of 918 matches played in the 2015/2016, 2016/2017 and 2017/2018 seasons in the Turkish Super League; data related to the variables of the number of goals thrown and defeated have been evaluated. Analysis of the matches played in the 2015/2016 and 2016/2017 seasons in the Turkish Super League has estimated the end of the 2017/2018 season of the league ranking. The data evaluated in the study was divided by a random method for training and testing. The team’s league rankings are modeled with numeric values between 0 (zero) and 1 (one). According to the analyses made with the developed YSA model, the Turkish Super League team ranking for many teams (test data set) was estimated with a accuracy rate of over 99%. In the Super League of Turkey, the team’s ranking last season has been determined and the number of goals defeated has been directly affected. The prediction of the final team ranking in football by the machine learning method can allow clubs to determine transfer policies according to the places they are targeting in the final league ranking.
This study, artificial neural networks (ANN) model using the Turkey Super League season-ending ranking of teams according to the number of input variables thrown and the renewed goal was conducted to predict. Working under the Turkey Super League in the 2015/2016, 2016/2017 and 2017/2018 a total of 918 matches played in the season; The data of the number of goals scored and defeated were evaluated. In the Turkey Super League, it was determined that seasonal data for 2015/2016 and 2016/2017 were as input variables, and seasonal data for 2017/2018 were output variables. The data analyzed in the study were separated randomly for training and testing purposes. The league order of the teams was modeled with numerical values between 0 (zero) and 1 (one). According to the results of the analysis conducted through the ANN model, the end-of-season team order in the Turkey Super League was estimated at high accuracy for several teams (above 99%) in the test dataset. Turkey Super League at the end of the season the team ranking is determined that directly affect the number of discarded and renewed goals. Estimating the end-of-season team ranking in football with the machine learning method can enable clubs to set transfer policies according to their destination in the end-of-season league ranking.
Journal Type : Ulusal
Article Detail 
Similar Articles
Download PDF
Journal Information
Listen Paper
Cite
Print This Page
Share
Mendeley
