適地商務服務之社群計算機制

Abstract

隨著社群媒體，基於適地性服務和行動裝置的出現，衍生出以社群網路服務結合手機平台和定位服務形式的行動商務新模式。本研究分別從消費者、供應商兩個面向進行探討，分別提出以行動社群網路為基礎的廣告代理人機制、廣告推薦機制以及行動社群支援決策機制，以期能有助於改善適地性商務的運作。 首先，如打卡、按讚等社群資訊以及如地點、時間、天氣等環境感知資訊，有助於推斷使用者對感興趣地點以及商店選擇的傾向，可藉此創造出適地性服務的商機。研究中結合個人的喜好分析、社會影響力分析和環境適合度分析，提出具環境感知的廣告推薦機制，可以成功地進行具有效益的廣告推薦，加快適地性商務的產生。 其次，考慮個人偏好、提供服務的所在地、行動手機用戶的習慣性移動軌跡以及代言人的影響力的因素，設計以具影響力的代理人來推播行動廣告機制來加強基於適地性廣告的有效性。 最後，由於網路社群上存在著眾多知識淵博的使用者，其意見足以作為使用者進行決策時的輔助。研究中所提出之行動社群支援決策機制，可以協助使用者將社群的知識力量化為己用，並考慮適地性達到即時決策支援之目的。透過此機制，可以成功地建議行動用戶下一個前往的興趣點。

With the emergence of social media, location-based services (LBS) and mobile devices, a new paradigm of mobile commerce can be envisioned. By taking into account the perspectives of customer and vendor, this thesis proposes several mechanisms (including the social context-aware advertisement recommendation mechanism, the social target advertising mechanism, and the social appraisal decision support mechanism) to enhance the location-based commerce. First, the social information (check-in, “like”, and so on) and contextual information (location, time, weather, and so on) are largely beneficial for inferring a user’s choice of place/store and can create significant opportunities for LBS. In this thesis, an advertisement recommendation mechanism is proposed that considers the methodologies of personal preference analysis, social influence analysis, and contextual fitness analysis, and can successfully leverage social and contextual information to effectively recommend LBS and expedite l-commerce. Secondly, a social endorsing advertising mechanism, which considers the factors of preference, location of service, the moving trajectories of mobile users, and the influential power of endorsers, is proposed to enhance the effectiveness of location-based advertising. Thirdly, social appraisal mechanism based on social decisions, incorporating factors of personal preference, social influence, geography, and consensus decisions, is proposed for mobile users. The proposed mechanism can successfully suggest POI for the mobile user to visit next.