以資料探勘技術建立病患就醫導引－－以胃腸科病患為例

Establishment of the Patient Guide Using Data Mining Techniques

由於全民健康保險的開辦實施，提供了人民生活上的醫療保障，但不可諱言的，由於醫療資源的浪費，也造成了政府財政上極大的負擔。究其原因，除了健保體制本身的問題外，國人未能具備正確的就醫觀念也是其中一項重大的因素。由於一般民眾並未具備足夠的醫療常識，因此往往在有了某些自覺症狀時，無法判斷是何種疾病所引起，以及是否該即時就醫，因而延誤病情的治療。這不僅將增加日後治療的困難，甚至可能危及生命，同時也讓社會增加了額外的醫療成本。資料探勘是近幾年在資料庫應用方面相當熱門的技術，並且廣泛的應用在各種領域。在本篇研究中，我們將資料探勘技術應用在醫療領域，期望能對上述問題有所改善。在探勘過程中，利用高效率之關聯規則探勘演算法(QDT)來歸納出各疾病與其可能症狀的高頻項目組，再以決策樹(Decision Tree)分類方法推論症狀與疾病兩者之間的關係，讓患者可藉由本研究的結果來瞭解其可能發生的疾病。本研究結果除了可作為一般民眾就醫時的參考資訊外，亦可提供醫師作為診斷時的參考依據，並期望能對維護全民健康有所貢獻。

The establishment of the National Health Insurance provides the medical protection for the people. However, it is also a heavy financial burden for the government because of the waste of the medical resources. There are many reasons to lead the waste of the medical resources. One of them is that most patients do not have enough medical senses. When patients have some symptoms, they cannot identify what disease they have, and they do not know which department they should consult Therefore, they often consult a wrong doctor and delay the treatment. It causes more difficulty in curing the disease and even brings forth the death of patients. Furthermore, it increases the additional social medical resources. Data mining techniques has been successfully applied in various fields. In this paper, we use the data mining techniques to process the medical data and establish the patient guide that can help patients to identify what disease they have. We employ both association rule and classification techniques for the most part of the mining process. First, Quick Decomposition Table algorithm is utilized to induce the large item sets between symptoms and diseases. Secondly, C4.5 algorithm that is one of the most popular decision tree techniques is adopted to classify the diseases what symptoms lead to. Finally, we can establish the patient guide to let patients know which diseases they may have. The patient guide can be a guide for the patients to consult the right doctor as well as a reference for the doctor to cure diseases. Moreover, we expect this research results can be helpful to human health.