利用適地性社群網路建構情境式廣告推薦機制

Abstract

適地性社群網路的興起，創造了更具黏著度的使用者及豐富的「社群內容」，而「行動廣告」則是社群網路平台視為未來主要的收益來源；傳統的行動廣告，除呈現形式有限外，派送的廣告內容與目標使用者之間的關聯性不佳，皆使得其效益被受質疑。以LBS（Location Based Services）廣告應用為例，當使用者依其位置查詢鄰近的廣告優惠時，派送服務通常僅依據「地理範圍」來派送廣告，過多「不適切」的資訊往往造成使用者的困擾。因此為更精確的進行廣告推播，本研究提出個人化的廣告推薦機制，除透過分析使用者興趣偏好外，同時以地理為基礎加入社群動態分析及情境感知資訊，如時間、天氣等因素，進而推薦給使用者個人化、適時、適地的廣告訊息，讓使用者有更強烈的意願到實體商店消費。

The increase usage of location base social network helped created a highly relevant user experiences and enriched "Social Media Content". With the rise of social network platforms, "Mobile Advertisement" has became the main source of revenues. Traditional mobile advertisement is not only limited in forms, but also in relevancy to the targeted viewers. At thus, its efficacy becomes questionable. With LBS (Location Based Services) advertisement, when users search for nearby sales and discounts, the services can deliver location based ad to the users; extra "Unneeded Information" can only cause users' agitations. To further the precision of advertisement delivery, this study will focus on propositions for advertisement personalization, and related recommendation mechanisms. Besides analyses of users' interests, location based info and social network info such as time, weather, etc are also added for personalized recommendations with timely and localized delivery with the intent of increasing users' motivations and willingness for in store purchases.