ΕΦΑΡΜΟΓΗ ΜΕΘΟΔΩΝ ΜΗΧΑΝΙΚΗΣ ΜΑΘΗΣΗΣ ΣΕ ΔΙΑΓΩΝΙΣΜΟΥΣ ΕΙΚΟΝΙΚΩΝ ΑΓΩΝΩΝ(FANTASY,GAMES) (Master thesis)
Χατζηγεωργίου, Κλεομένης
In this thesis we developed a machine learning model that predicts fantasy performance of NBA players, based on their past performance. We tried three different regression approaches, Bayesian Regression, Linear Regression and Support Vector Regression. We also developed an algorithm that uses dynamic programming to generate a daily fantasy sports lineup based on our model’s predictions.
We ended up using Bayesian Regression for our model due to better produced results compared to the other approaches. Next we generated lineups for random contests of NBA 2016-2017 regular season. We also compared our results with the predictions of websites that offer similar services. Through our tests we managed to produce several lineups that could potentially return profit in the long term and at the very least offer a tool that can be used by daily fantasy players to get an edge over their competition.
| Institution and School/Department of submitter: | Σχολή Τεχνολογικών Εφαρμογών / Τμήμα Μηχανικών Πληροφορικής |
| Keywords: | Μηχανική Μάθηση, Παλινδρόμηση, Φαντασιακά Αθλήματα |
| URI: | http://195.251.240.227/jspui/handle/123456789/11226 |
| Table of contents: | Table of Contents 1 Introduction 1 1.1 Data Analysis in Sports 1 1.2 Fantasy Sports 1 1.2.1 Industry Background 2 1.2.2 Scoring and Format 3 1.2.3 Fantasy Contests 4 1.3 Project Goal 5 1.3.1 Contribution 5 2 Related Works 7 3 Theoretical Background 8 3.1 Using machine Learning in DFS 8 3.2 Machine Learning 8 3.3 Regression 10 3.3.1 Linear Regression 10 3.2.2 Support Verctor Machines (SVM) 11 3.2.3 Random Forests 13 3.3 Python 13 3.3.1 Scikit-learn 14 4 Development 15 4.1 Creating the dataset 15 4.2 Predicted minutes 16 4.3 Developing a prediction model 18 5.3 Lineup Optimization 21 5 Evaluation 24 5.1 Assessment parameters 24 5.2 Tests 25 6 Technical details 42 6.1 Web Scrapper 42 6.2 Lineup Optimizer 44 7 Conclusion 49 7.1 Outline and conclusions 49 7.2 Future work 50 8 Annex….. 49 8.1 Web Scrapper 51 8.2 Prediction model 52 8.3 Lineup Optimizer 54 9 References 60 |
| Appears in Collections: | Μεταπτυχιακές Διατριβές |
Files in This Item:
| File | Description | Size | Format | |
|---|---|---|---|---|
| CHATZIGEORGIOY.pdf | Μεταπτυχιακή | 12.59 MB | Adobe PDF | View/Open |
Please use this identifier to cite or link to this item:
This item is a favorite for 0 people.
http://195.251.240.227/jspui/handle/123456789/11226
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.
