Publication:

植基於資料挖掘技術的體適能評級模式建構之研究

資料挖掘(Data mining) 是最近被提出來的一種資料處理的技術，它結合了統計理論及機器學習技術。本文嘗試使用資料探勘技術以建構一女性青少年的體適能評估模式，藉以科學化的區別青少年的體適能之等級。由於本模式係針對女性青少年所建構，因此樣本選用台中技術學院五專部的新生女學生。模式建構之前，首先選擇最能代表體適能的各項測驗指標，此項測驗指標的選擇是以教育部所公佈的體適能評鑑項目為依據(包含：身體組成、柔軟度、腹肌肌力或肌耐力、瞬發力、心肺耐力等各項測驗)。

模式建構的程序如下：首先，收集樣本，續而依據各學生的體適能資料，將所有學生進行群組分析(形成不同類別體適能的群組)。再將各體適能群組分別利用區別分析、神經網路、支援向量機等三項技術建構體適能評估模式，作為青少女體適能評級分類的基礎，最後再以用此體適能評估模式，以協助學生了解自己的體適能，藉以提升國民健康程度，此亦為本文最大之目的。

Abstract

Data Mining have been recently introduced as a new technique for data processing field. This research is to construct a fitness model of female teenagers, which is based on the data mining techniques. By using this model, we can classify the physical fitness scales of teenagers. The samples for this model are 15 years old National Taichung Institute of Technology freshmen. Before constructing this model, we have chosen the most reputable exercise index of physical fitness. The selected criteria are based on the data of The Department of Education. The assessment of physical fitness includes body composition, flexibility, abdominal strength or muscle endurance, anaerobic power and cardiorespiratory endurance.

The fitness models have been constructed as follow: First, collecting the sample and using the Clustering method to form the different cluster of fitness. Second, constructing the fitness models based on the Discriminant Analysis、Neural Netwoks、Support Vector Machine. Finally, utilizing these models to guide students not only to understand their physical fitness scales, but also to improve their health.