Information

  • Publication Type: Bachelor Thesis
  • Workgroup(s)/Project(s):
  • Date: June 2022
  • Date (to): 22. June 2022
  • Matrikelnummer: 11822449
  • First Supervisor: Eduard Gröller

Abstract

This Paper is centered around the topic of predicting the winner of a Tennis-Match on the ATP-Tour. Firstly there is a short introduction about the relevance and research of Machine Learning. At first the topic will be discussed broadly, then more and more narrowly to match the goal of the thesis. Next a research for important literature will be carried out, to present the state of the art in the field of Machine Learning based predictions, in science and on the market. Prediction algorithms will be discussed, with an emphasis on outcome prediction in different sports. As an important and high revenue usecase of prediction, the methods of betting companies and bookmakers also get a chapter to be discussed. Same applies to the visualization of data in conjunction to Machine Learning, related Frameworks and Libraries and lastly also future challenges of applying Machine Learning In the practical part of the thesis, an application will be talked about, that compares the performance of different neural Networks that try to predict the winner of a Tennis- Match. To accomplish that, at first a dataset with over 200.000 entries and more than 200 variables was split into four chunks to be the input for one neural Network each. Those neural Networks were made and trained in Tensorflow and their accuracy was tested via picking three players and trying to predict their matches. The results were then plotted into a Scatterplot. With this methodology it is possible to get information about what variables, that the public has access to before the match starts, are how important for the outcome of a match. Furthermore it would be feasible to find out, what set of variables is best suited to make predictions about the winner and also if there are differences in the predictions of the three players.

Additional Files and Images

Additional images and videos

Additional files

Weblinks

No further information available.

BibTeX

@bachelorsthesis{Leutschacher_2022,
  title =      "Visual Analysis of the Prediction of ATP-Matches",
  author =     "Manuel Leutschacher",
  year =       "2022",
  abstract =   "This Paper is centered around the topic of predicting the
               winner of a Tennis-Match on the ATP-Tour. Firstly there is a
               short introduction about the relevance and research of
               Machine Learning. At first the topic will be discussed
               broadly, then more and more narrowly to match the goal of
               the thesis. Next a research for important literature will be
               carried out, to present the state of the art in the field of
               Machine Learning based predictions, in science and on the
               market. Prediction algorithms will be discussed, with an
               emphasis on outcome prediction in different sports. As an
               important and high revenue usecase of prediction, the
               methods of betting companies and bookmakers also get a
               chapter to be discussed. Same applies to the visualization
               of data in conjunction to Machine Learning, related
               Frameworks and Libraries and lastly also future challenges
               of applying Machine Learning In the practical part of the
               thesis, an application will be talked about, that compares
               the performance of different neural Networks that try to
               predict the winner of a Tennis- Match. To accomplish that,
               at first a dataset with over 200.000 entries and more than
               200 variables was split into four chunks to be the input for
               one neural Network each. Those neural Networks were made and
               trained in Tensorflow and their accuracy was tested via
               picking three players and trying to predict their matches.
               The results were then plotted into a Scatterplot. With this
               methodology it is possible to get information about what
               variables, that the public has access to before the match
               starts, are how important for the outcome of a match.
               Furthermore it would be feasible to find out, what set of
               variables is best suited to make predictions about the
               winner and also if there are differences in the predictions
               of the three players. ",
  month =      jun,
  address =    "Favoritenstrasse 9-11/E193-02, A-1040 Vienna, Austria",
  school =     "Research Unit of Computer Graphics, Institute of Visual
               Computing and Human-Centered Technology, Faculty of
               Informatics, TU Wien ",
  URL =        "https://www.cg.tuwien.ac.at/research/publications/2022/Leutschacher_2022/",
}