Incorporating domain knowledge in machine learning for soccer outcome prediction

The task of the 2017 Soccer Prediction Challenge was to use machine learning to predict the outcome of future soccer matches based on a data set describing the match outcomes of 216,743 past soccer matches. One of the goals of the Challenge was to gauge where the limits of predictability lie with this type of commonly available data. Another goal was to pose a real-world machine learning challenge with a fixed time line, involving the prediction of real future events. Here, we present two novel ideas for integrating soccer domain knowledge into the modeling process. Based on these ideas, we developed two new feature engineering methods for match outcome prediction, which we denote as recency feature extraction and rating feature learning. Using these methods, we constructed two learning sets from the Challenge data. The top-ranking model of the 2017 Soccer Prediction Challenge was our k-nearest neighbor model trained on the rating feature learning set. In further experiments, we could slightly improve on this performance with an ensemble of extreme gradient boosted trees (XGBoost). Our study suggests that a key factor in soccer match outcome prediction lies in the successful incorporation of domain knowledge into the machine learning modeling process.

     

Learning to predict soccer results from relational data with gradient boosted trees

We describe our winning solution to the 2017’s Soccer Prediction Challenge organized in conjunction with the MLJ’s special issue on Machine Learning for Soccer. The goal of the challenge was to predict outcomes of future matches within a selected time-frame from different leagues over the world. A dataset of over 200,000 past match outcomes was provided to the contestants. We experimented with both relational and feature-based methods to learn predictive models from the provided data. We employed relevant latent variables computable from the data, namely so called pi-ratings and also a rating based on the PageRank method. A method based on manually constructed features and the gradient boosted tree algorithm performed best on both the validation set and the challenge test set. We also discuss the validity of the assumption that probability predictions on the three ordinal match outcomes should be monotone, underlying the RPS measure of prediction quality.

     

Soccer server: A tool for research on multiagent systems

This article describes Soccer Server, a simulator of the game of soccer designed as a benchmark for evaluating multiagent systems and cooperative algorithms. In real life, successful soccer teams require many qualities, such as basic ball control skills, the ability to carry out strategies, and teamwork. We believe that simulating such behaviors is a significant challenge for computer science, artificial intelligence, and robotics technologies. It is to promote the development of such technologies, and to help define a new standard problem for research, that we have developed Soccer Server. We demonstrate the potential of Soccer Server by reporting an experiment that uses the system to compare the performance of a neural network architecture and a decision tree algorithm at learning the selection of soccer play plans. Other researchers using Soccer Server to investigate the nature of cooperative behavior in a multiagent environment will have the chance to assess their progress at RoboCup-97, an international competition of robotic soccer to be held in conjunction with IJCAI-97. Soccer Server has been chosen as the official server for this contest.     

The importance of a match in a tournament

A quantitative measure of “match importance” is useful in a number of decision problems, for example: as a metric in tournament design; for selecting matches for broadcasting; for scheduling matches in a tournament; and for assigning referees. To date measures of match importance used in such analyses have been relatively naïve. We discuss a general measure that considers the effect of a particular match on the end of tournament position, given the results of all other matches, some played, some predicted. We use logistic regression to predict matches and Monte Carlo simulation to compute the match importance measure, and apply these to soccer matches in the English Football Association Premier League.     

Exploring polynomial classifier to predict match results in football championships

Football is the team sport that mostly attracts great mass audience. Because of the detailed information about all football matches of championships over almost a century, matches build a huge and valuable database to test prediction of matches results. The problem of modeling football data has become increasingly popular in the last years and learning machine have been used to predict football matches results in many studies. Our present work brings a new approach to predict matches results of championships. This approach investigates data of matches in order to predict the results, which are win, draw and defeat. The investigated groups were different type of combinations of two by two pairs, win-draw, win-defeat and draw-defeat, of the possible matches results of each championship. In this study we employed the features obtained by scouts during a football match. The proposed system applies a polynomial algorithm to analyse and define matches results. Some machine-learning algorithms were compared with our approach, which includes experiments with information obtained from the football championships. The association between polynomial algorithm and machine learning techniques allowed a significant increase of the accuracy values. Our polynomial algorithm provided an accuracy superior to 96%, selecting the relevant features from the training and testing set.     

Layered approach to learning client behaviors in the robocup soccer server

In the past few years, multiagent systems (MAS) have emerged as an active subfield of artificial intelligence (AI). Because of the inherent complexity of MAS, there is much interest in using machine learning (ML) techniques to help build multiagent systems. Robotic soccer is a particularly good domain for studying MAS and multiagent learning. Our approach to using ML as a tool for building Soccer Server clients involves layering increasingly complex learned behaviors. In this article, we describe two levels of learned behaviors. First, the clients learn a low-level individual skill that allows them to control the ball effectively. Then, using this learned skill, they learn a higher level skill that involves multiple players. For both skills, we describe the learning method in detail and report on our extensive empirical testing. We also verify empirically that the learned skills are applicable to game situations.     

A Lagrangian approach for minimum cost single round robin tournaments

Single round robin tournaments are a well known class of sports leagues schedules. We consider leagues with a set T of n teams where n is even. Costs are associated to each match and period. Since matches are carried out at one of the both opponents venues matches may be forbidden in certain periods due to unavailability of stadiums. The goal is to find a minimum cost tournament among those having no match scheduled infeasible. We employ a Lagrangian relaxation approach in order to obtain tight lower bounds. Moreover, we develop a cost-oriented repair mechanism yielding a feasible tournament schedule to each solution of the relaxed problem.     

Forecasting market prices in a supply chain game

Predicting the uncertain and dynamic future of market conditions on the supply chain, as reflected in prices, is an essential component of effective operational decision-making. We present and evaluate methods used by our agent, Deep Maize, to forecast market prices in the trading agent competition supply chain management game (TAC/SCM). We employ a variety of machine learning and representational techniques to exploit as many types of information as possible, integrating well-known methods in novel ways. We evaluate these techniques through controlled experiments as well as performance in both the main TAC/SCM tournament and supplementary Prediction Challenge. Our prediction methods demonstrate strong performance in controlled experiments and achieved the best overall score in the Prediction Challenge.     

Modeling outcomes of soccer matches

We compare various extensions of the Bradley–Terry model and a hierarchical Poisson log-linear model in terms of their performance in predicting the outcome of soccer matches (win, draw, or loss). The parameters of the Bradley–Terry extensions are estimated by maximizing the log-likelihood, or an appropriately penalized version of it, while the posterior densities of the parameters of the hierarchical Poisson log-linear model are approximated using integrated nested Laplace approximations. The prediction performance of the various modeling approaches is assessed using a novel, context-specific framework for temporal validation that is found to deliver accurate estimates of the test error. The direct modeling of outcomes via the various Bradley–Terry extensions and the modeling of match scores using the hierarchical Poisson log-linear model demonstrate similar behavior in terms of predictive performance.

     

集控式足球机器人系统关键技术研究

足球机器人比赛是机器人研究的一个新热点,它为人工智能理论和算法的研究提供了一个实验平台,其研究的领域涵盖了人工智能、自动控制、机器人视觉、无线通信、机器学习和多智能体合作与协调等。集控式足球机器人系统通常可以划分为4个子系统,即视觉、决策、通信和车型机器人。结合研究经验,介绍了集控式足球机器人各个子系统的关键技术。     

强化学习在机器人足球比赛中的应用

机器人足球比赛是一个有趣并且复杂的新兴的人工智能研究领域 ,它是一个典型的多智能体系统。采用强化学习方法研究了机器人足球比赛中的足球机器人的动作选择问题 ,扩展了单个Agent的强化学习方法 ,提出了基于多Agents的强化学习方法 ,最后给出了实验结果。     

基于人工神经网络的多机器人协作学习研究

机器人足球比赛是一个有趣并且复杂的新兴的人工智能研究领域，它是一个典型的多智能体系统。文中主要研究机器人足球比赛中的协作行为的学习问题，采用人工神经网络算法实现了两个足球机器人的传球学习，实验结果表明了该方法的有效性。最后讨论了对BP算法的诸多改进方法。     

基于赛场态势的群体智能决策技术

复杂环境下态势的正确评估是开发高水平决策系统所必需解决的关键问题。为了对机器人足球赛场态势进行精确的评估,结合专家经验和对实际比赛的分析,提炼影响赛场态势的几个关键因素,在此基础上提出一个全新的态势决策模型。该模型中机器人集合构成一个决策群体,智能群体决策的结果使得足球机器人系统具有很强的战斗力,其有效性已在实验及比赛中得到验证。     

Optimizing player’s formations for corner-kick situations in RoboCup soccer 2D simulation

In the domain of the Soccer simulation 2D league of the RoboCup, appropriate player positioning against the opponent team formation is an important factor of soccer team performance. In this work, we propose to use a meta-heuristic algorithm called the firefly algorithm to optimize player positioning. We used sequential Bayesian estimation as well as parallelization to reduce the necessary number of time-consuming simulated soccer matches. As a first trial of our system, we optimized the corner-kick formation. Preliminary results in optimizing the corner-kick formation are not advantageous over the previous handmade formation due to the difficulty in tuning the meta-heuristic algorithm parameters. However, it is also shown that the proposed system is effective in handling a high load of simulations over the span of weeks and therefore is promising to be usable to optimize player positioning.     

足球视频中的目标检测与跟踪

提出了一种新颖的目标检测与跟踪算法来检测和跟踪足球运动中的球员。与普通的利用颜色分割的方法不同,考虑到足球视频中非目标的像素大体上都是单一的绿色这个特点,结合颜色的统计信息和像素的边缘特性来得到更完美的检测效果;接着利用灰度图像中的统计信息,轻松地完成球员队属的辨别;最后根据重叠面积提出一种简单的视频目标跟踪方案,结合图像匹配,解决运动中的遮挡问题。     

类人足球机器人决策系统的设计

类人机器人足球比赛是机器人足球比赛的最高赛事.类人足球机器人的决策系统是基于独立视觉的自主决策系统,很大程度上决定着比赛的胜败.介绍了自主研发的类人足球机器人决策系统的架构及实现方法,并在此基础上运用有限状态机理论,对单个机器人的自主进攻策略进行了详细分析和研究,真实环境中的实验及比赛结果证明了其有效性.该决策系统的设计及研究工作对基于自主决策的多智能体协作以及服务性机器人决策系统的研究都具有重要的价值.     

基于机器学习的风险预测方法研究

预测是用科学的方法和手段时事物的发展趋势和未来状态进行估量的技术.为了弥补传统方法和技术的不足,各种机器学习技术越来越多地应用于预测的研究中.讨论了在风险预测这一特定领域,应用基于案例的推理(CBR,Case based Reasoning)、支持向量机(SVM,Support Vectot Machine)以及人工神经网络(ANN,Artificial Neural Network)等机器学习方法来进行预测的技术.同时,以我们的工作为基础,详细论述了在信贷风险预测和工程评标中基于机器学习预测模型的应用.     

分层技术在机器人足球中的应用研究

多智能体学习是多智能体系统和机器学习等研究领域的交叉。由于多智能体系统的复杂性,研究者们采用基于分层的机器学习的方法来解决。设计了一个基于分层技术的决策子系统,在此基础上,使用机器学习的方法对分层之后的决策学习和技能学习进行了研究实现。最后,指出了当前机器人足球中分层学习技术存在的不足,为以后机器人足球研究又增加了一个新的研究课题。     

Using a SAT-solver to schedule sports leagues

Tournament schedules of sports leagues have to satisfy several types of constraints such as stadium unavailability, fixed matches, forbidden matches, minimum number of breaks. Usually, there is no schedule satisfying all given constraints and, hence, some of the constraints are considered as ‘soft’ ones. There are various models appropriately describing the environment of sport leagues. Only heuristic methods are known from the literature for solving instances of real life dimensions. We consider here a model which satisfies the demands of many sports leagues. We solve our model by reduction to series of instances of the propositional satisfiability problem and adaption of a satisfiability solver for these specific instances. We test our method on two real life examples and solve the problem optimally within our model in each case. Our solver shows good computational results also on generated test instances, which are motivated by real life requirements. It can be easily extended to meet the demands of other sports leagues.     

A general‐purpose task execution framework for manipulation mission of the 2017 Mohamed Bin Zayed International Robotics Challenge

This paper presents the hardware and software of our team's EurecarBot for Challenge 2 in the 2017 Mohamed Bin Zayed International Robotics Challenge (MBZIRC). Fully automating our robots actions in a real environment required many component technologies for manipulation and vision processing. To perform the complex robotic missions, we developed a task execution framework, which provides a high‐level interface to specify the given tasks. In this study, we focus on the valve operation problem, which was the hardest part of the competition. We also discuss how we overcame the various problems caused by differences between the experimental and the actual competition environments. EurecarBot completed the valve operation mission perfectly in the MBZIRC Grand Challenge and ranked fourth in Challenge 2 and fifth in the Grand Challenge.