以個人化建構微時刻推薦系統的互動機制

A personalized Interactive Mechanism Framework for Micro-moment Recommender System

李紹威(Lee Shao-Wei)

指導教授 ： 林怡伶

國立政治大學/商學院/資訊管理學系/碩士(2021年)

https://doi.org/10.6814/NCCU202101336

摘要

微時刻概念的出現凸現了情境對人們造成的影響，而推薦系統應該要順應這樣的趨勢做出改變。為了搜集到足夠的情境資料，微時刻推薦系統必須要有有效的互動機制，讓使用者和系統之間可以方便的互動。本研究採用了支援自治和本體的設計原理，混合不同種類的個人化去設計了四種互動機制，並且將他們實作在一個微時刻推薦應用程式中。本研究的目的是想了解哪一種互動機制最適合微時刻推薦系統的互動機制，根據我們採用的設計原理和微時刻推薦系統的特性，我們認為愈能讓使用者掌控系統和花費較少心力的設計應該會較為適合。我們藉由為期兩週的受測者間實驗去驗證我們的假設。在實驗中我們讓受測者實際使用我們的應用程式，並收集他們的回饋和使用時的紀錄。我們發現在不同的互動機制中存在控制感受的差異，以採用使用者發起和使用者與系統共同發起的個人化的互動機制較高，而且額外的控制不會讓受測者花費多餘的心力。因此我們認為這兩種設計較適合微時刻推薦系統的互動機制。

關鍵字

微時刻推薦系統 ； 個人化 ； 動機賦能 ； 互動機制

並列摘要

The emergence of the micro-moment concept highlights the influence of context, and the recommender system should be adjusted according to this trend. In order to collect enough contextual information, the micro-moment recommender system (MMRS) have an effective interactive mechanism that allows users to easily interact with the system. This study adopts the design principle of supporting autonomy and promoting the creation and expression of self-identity, mixes different types of personalization to design four types of interactive mechanisms, and implements them in a micro-moment recommender app. The purpose of this study is to understand which interactive mechanism is the most suitable for MMRS. Based on the design principles we adopted and the characteristics of MMRS, we believe that the design that allows users to have more control over the system and uses less effort should be more suitable for supporting micro-moment needs. We tested our hypothesis by a two-week between-subject field study. In the field study, the participants use our app and provide their feedback. We found that there is a difference in perceived active control among different interactive mechanisms, with user-initiated personalized intention and mix-initiated personalized intention personalization mechanisms having higher perceived active control, and the additional control does not cost the participants extra effort. Therefore, we believe that these two designs are more suitable for the MMRS interactive mechanism.

並列關鍵字

Micro-moment recommender system ； Personalization ； motivational affordance ； Interactive mechanism

參考文獻

Abowd, G. D., Dey, A. K., Brown, P. J., Davies, N., Smith, M., and Steggles, P. 1999. “Towards a Better Understanding of Context and Context-Awareness,” in Lecture Notes in Computer Science (Including Subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics).

Adomavicius, G., Mobasher, B., Ricci, F., and Tuzhilin, A. 2011. Context-Aware Recommender Systems.

Alsalemi, A., Sardianos, C., Bensaali, F., Varlamis, I., Amira, A., and Dimitrakopoulos, G. 2019. “The Role of Micro-Moments: A Survey of Habitual Behavior Change and Recommender Systems for Energy Saving,” IEEE Systems Journal (13:3), IEEE, pp. 3376–3387.

Baltrunas, L., Ludwig, B., Peer, S., and Ricci, F. 2013. “Context Relevance Assessment and Exploitation in Mobile Recommender Systems,” Personal and Ubiquitous Computing (16:5), pp. 507–526.

Barkhuus, L., and Dey, A. 2003. “Is Context-Aware Computing Taking Control Away from the User? Three Levels of Interactivity Examined.”