建設公司問卷資料探勘系統設計

Abstract

建設公司往往透過問卷了解客戶對產品之需求，人員彙整問卷資料後，由於對統計軟體的熟悉度有限，通常僅以敘述性統計量做為決策輔助的依據，而忽略各屬性之關聯分析。分析各屬性之關聯往往能協助調整公司之策略及作為，為了改善問卷資料中不同屬性間過於獨立之不足，本研究以資料探勘技術為基礎，以建設公司人員為目標使用族群，利用開放式之R軟體及Shiny套件為工具，實作一套互動式系統，以提供不具資訊、統計背景的使用者實行資料探勘中的關聯分析。

Construction company’s collected questionnaire data is usually only descriptive statistics as a basis for decision support. As a result, it may ignore the correlation between the hidden attribute. In order to improve the shortage which is questionnaire data’s different attributes are too independent, we can use data mining techniques to enter. However general construction companies do not have this expertise. They have entrusted to other vendors. To solve this problem, this study uses R software and Shiny package as a tool and implements an interactive system. The proposed system allows the user without strong information technology and statistics background to be able to perform data mining analysis.