Mining Olympic Sailing Boat Telemetry Data

目前的机载技术意味着每个采样点可以获得关于帆船状态的大约十个方面的数据。这些包括通过GPS定位，舵角，风向角，风速，3轴加速度计等。此后有可能推导出其他的。视频和天气信息通常也是可用的。到目前为止，基于这些数据的调整，如果使用的话，都依赖于教练和/或运动员的解释。有人提出，数据可能包含关于变量和船速的先前未知的信息，因此应该探索数据挖掘。因此，数据挖掘任务是数值预测之一，并且将探索许多众所周知的方法。此后，将采用知识发现方法，特别是在人工神经网络的情况下，明确的特征选择和规则提取。

Current on-board technology means it is possible to obtain data with respect to around ten aspects of a sailing boat's state per sample point. These include position via GPS, rudder angle, wind angle, wind speed, 3-axis accelerometer, etc. It is possible to derive others thereafter. Video and weather information is also typically available. Thus far adjustments based on such data, if it is utilised at all, have relied upon the interpretation of the coach and/or athletes. It is proposed that the data may contain previously unknown information regarding the variables and boat speed and hence that data mining should be explored. The data mining task is thus one of numerical prediction and a number of well-known approaches will be explored. Thereafter approaches to knowledge discovery will be employed, particularly explicit feature selection and rule extraction in the case of artificial neural networks.