劉彥辰

Yan-ChenLiu

碩士論文（2012）

虛擬社群之潛在顧客搜索機制研發-以食品業應用為例

Development of a Mechanism of Potential Customer Searching from Virtual Communities: Food Industry as an application

關鍵字 Keywords

消費價值模式、向量空間模型、潛在顧客、潛在語意分析、倒傳遞類神經網路

Consumption Value Theory、Virtual Community、Vector Space Mode、Latent Semantic Analysis、Back Propagation Neural Network

摘要

過去研究顯示，開發一個顧客所消耗的成本為保留顧客的五倍。若能降低開發顧客之成本，將有助於企業提升利潤。現有之潛在顧客獲取方法多以市場調查或資料探勘為主，前者經常花費許多人力及時間；後者因無法取得競爭對手資料，導致無法了解整體市場趨勢、消費者偏好的變化。因此，如何有效率地獲取潛在顧客為當前企業重要之議題。

本研究以消費價值理論(Consumption Value Theory)為基礎，發展一自動於虛擬社群中尋找潛在顧客之機制，藉由自動化之方法獲取潛在顧客，以期降低企業所投入之人力及資源。由於消費價值理論難以量化為自動化處理之依據，本研究遂進一步提出以食品業為基之產品一般屬性模型(Product Common Attribute model，PCA)與消費價值理論對應。透過問卷調查，檢驗該理論與一般屬性模型之關聯性，進一步將所獲得之各項路徑係數建立消費價值與產品一般屬性對應之權重矩陣。本研究將虛擬社群中之網路文章，以潛在語意分析(Latent Semantic Analysis，LSA)以及倒傳遞類神經網路(Back Propagation Neural Network，BPNN) 輔以上述權重矩陣進行計算，作為識別潛在顧客之方法，篩選出可能為潛在顧客之網路使用者。

Abstract

The cost for acquiring new customers is more than five times the cost for satisfying and retaining current customers. Thus, reducing costs of acquiring new customers is an important issue for maximizing enterprise profit.

This study aims to develop a potential customer searching mechanism based on consumption value theory. To acquiring the weight of consumption value theory in application domain , This study has proposed a product common attribute model to fit such theory by questionnaire survey. Depending on acquired weight of consumption value theory, we then develop a potential customer identification method to directly locate customers that provide further information to enterprise automatically, and enable to acquire customers without wasting human resources and money.