Tennis Match

Tennis is one of the most popular sports played in the world. Many researchers have worked in the fields of forecasting the outcome of tennis matches using past statistical data records. This paper mainly investigates the comparison between three different classifiers namely decision tree, multilayer perceptron and support vector machine. The research study aims to predict the result of tennis matches using eight UCI databases of grand slam tennis tournaments and evaluate the classification performance using various measures such as the root-mean-square error, accuracy, false positive rate, true positive rate, kappa statistic, recall, precision, and f-measure. All these performance measures confirm the supremacy of the decision tree classification

The tennis prediction model is developed to evaluate the chance of winning match that the players will face. When a game is played, the result depends on many factors including the playing environment, player’s skill and past match results. Many approaches such as statistical data evaluation have been used so far. But predicting the theoretical outcome of tennis matches is a challenging task and has been a keen interest for many researchers. Indeed, enough scope is there for making significant improvement in the quality of prediction and the interpretation of results. The present research study basically aims to predict the outcome of a tennis singles match using past match records of the grand slam

The artificial neural network [7] [8] [9] is a computational model inspired by human central nervous systems used to estimate or approximate functions that depends on a large number of unknown input data. In this model, we use multilayer perceptron (MLP) [10] [11] which is a feed-forward artificial neural network model that maps sets of input data elements considered as individual nodes on to a set of appropriate output data elements. An MLP contains multiple layers of nodes in a directed graph in which each of the different layers fully connected to the next layer. Also, each node is a neuron (or individual processing element) with a nonlinear activation function except for the input nodes. It uses a supervised learning method called backpropagation for training the neural network. Support vector machine (SVM) [12] is a supervised learning model that can analyze data and recognize patterns used for classification and regression analysis. The decision tree [13] [14] learning approach considers decision tree as a prediction model that is a tree with internal nodes as each decision and leaf nodes as the result of the decisions made. This approach can be used to analyze our result as each path from the root to leaf node represents a solution for our