應用資料探採於我國西藥行銷之研究

Abstract

應用資料探採於我國西藥行銷之研究 研究生：廖雅郁 指導教授：陳光華 教授 論 文 摘 要 「資料探採」(Data Mining)是結合資訊工程、分析技術及產業專業知識的綜合技術，可以協助終端使用者從大型資料庫擷取出有用的資訊並轉換為知識。而且資料探採在商業的應用涵蓋許多領域，如顧客獲利率分析、獲取顧客、市場區隔、目標行銷、資產風險管理、信用卡風險管理、交叉銷售等。因此本研究針對製藥市場應用資料探採的技術，分析製藥業者之客戶群，了解其客戶群之處方偏好型態進而根據分析之結果讓業者可以靈活運用各種行銷策略，具體地達到提昇營業額、改善獲利率、減少成本、增進客戶滿意度等績效目標。 本研究選擇製藥市場為研究對象是因為製藥的發展象徵一個國家的進步程度，國民所得越高的國家，製藥產業越顯發達，而隨著全民健保及相關藥政措施的陸續施行，國內藥品市場結構改變、市場秩序日趨完善，面對國內高齡化的社會，無疑也透露著一種藥品消費市場會大幅成長的訊息，因此找出市場動態脈絡對於業者制訂行銷策略將有極大的幫助。 本研究主要的目的為1.探討製藥市場概況。2.探討影響醫師處方型態的可能因素。3.歸納可能影響醫師處方型態的因素，進而建構樣本醫師族群的資料庫。4.應用資料探採技術分析，探討處方者「人口統計變數」、「資訊尋求變數」、「心理區隔變數」、「行為區隔變數」在處方型態分析結果規則中所扮演的角色，並且作為各個分析規則的特徵描述。5.依據上述的分析結果，提出行銷策略與建議，提供製藥業者選擇目標市場及制訂行銷策略之參考。 本研究以決策樹學習法進行分析，產生之結論如下： 1.以11種屬性區隔市場，其屬性分別如下:年齡、教育程度、職稱、醫院性質、醫院層級、醫院政策、藥廠銷售人員拜訪數目、認同國內藥廠的努力、認為同成分藥品來源皆相同、本身臨床使用經驗與最早使用新藥的時間。 2.建立13種implication sentences 描述醫師處方偏好之行為特色。

Study on application of data mining for pharmaceutical market in Taiwan Student:Ya-Yu Liao Advisor:Prof. Guang-Hwa Chen Abstract Data mining is a comprehensive technology which combines information technology, analysis technique and industry know-how. With this technology, user can discover modeling of hidden patterns in large volumes of data and transform such data into information and even knowledge. The business applications in data mining include profitability analysis, customer attraction, market segmentation, asset risk management, and cross selling. This study applies data mining techniques in pharmaceutical market. The pharmaceutical companies can use the results from prescription on customers’ behavior to expand their business performance. Data mining creates information assets that a company can leverage to achieve their strategic objectives. The formation of aging society in Taiwan became notable since 1994, and the national health policy has been carried out since 1995. Consequently, the pharmaceutical market in Taiwan has a significant growth out of increasing medical demand and medical care in recent years. With the rapid improvement of data processing technology, the pharmaceutical companies can fully exploit data mining techniques to discover market trends and thus to further adopt appropriate marketing strategies. Therefore, purposes of this study include (1) understanding the current domestic pharmaceutical market, (2) finding out possible factors which can influence prescription behavior of physicians, (3) constructing a database for physicians, (4) applying decision tree method to analyze database constructed above, and (5) providing appropriate strategies so that pharmaceutical companies can make proper uses to achieve their objectives. The conclusions of the study obtained are listed as follows. (1) The pharmaceutical market currently in Taiwan can be segmented by 11 attributes. (2) Prescription behavior of physicians is influenced by 13 implication sentences.