Artificial Neural Network (ANN) is one of the very popular artificial intelligent tools for prediction. There are three researchers called Amornchai Somboonphokkaphan, Suphakant Phimoltares, and Chidchanok Lursinsap. They applied ANN modeling to predict the winner of the tennis match. The name of the academic paper is “Tennis Winner Prediction based on Time-Series History with Neural Modeling” which is a conference paper in 2009.
(人工神经网络（ANN）是一种非常流行的人工智能工具预测。有三个研究人员称为 Amornchai Somboonphokkaphan，Suphakant Phimoltares和Chidchanok Lursinsap。他们采用人工神经网络模型来预测网球比赛的胜利者。这个名字的学术论文是“基于时间序列的神经建模历史对网球比赛的预测”，这是在2009年的学术会议文件。 )
They used a ANN network to build a tennis match model. They collected the statistical data of each player in the past few years until the day before prediction. They added an input parameter which is called the “court surface”. According to the “court surface”, it produces an effect to the individual statistic of the player.
(他们用人工神经网络的网络，以建立一个网球比赛模型。他们收集的每个球员在过去数年的统计数据，直到预测的前一天。他们将一个输入参数，它被称为“场地表面”。根据“场地表面”，它产生的效果，以个别的球员统计。)
The selected statistical features includes
1. Winning percentage on the first serve;
2. Winning percentage on the second serve;
3. Winning percentage on the return serve;
4. Winning percentage on the break point;
5. Winning percentage of played match; and
6. Total points win.
7. Hard Court or Clay Court or Grass Court
(选定的统计特征包括
1。第一发球局胜率;
2。在第二发球局胜率;
3。在返回发球局胜率;
4。在决胜局胜率;
5。全场比赛的胜率;及
6。总积分。
7。场地表面)
They used a few year data as training sets. They showed about 75% accuracy in the Australian Open 2003 Tournament in the training set. They also have a time series model to show 70% to 80% accuracy on 2007 -2008 data of Australian Open, French Open, Winbledon and US Open Tournaments.
(他们用数年的数据作为训练集。他们发现在训练集大约 75％的准确率在2003年澳大利亚网球公开赛赛。他们也有一个时间序列模型显示 70％至80％在2007 -2008数据准确性的澳大利亚公开赛，法国公开赛，美国公开赛和锦标赛 Winbledon。)
Reference:
Amornchai Somboonphokkaphan, Suphakant Phimoltares, and Chidchanok Lursinsap (2009), “Tennis Winner Prediction based on Time-Series History with Neural Modeling”, Proceedings of the International MultiConference of Engineers and Computer Scientists, Vol I, IMECS 2009, March 18-20, Hong Kong
