摘要
隨著社群網絡的蓬勃發展，電影公司開始通過各種社群平台宣傳電影，例如Instagram。因此，在本研究中，我們提出了一個研究框架推薦電影帳戶給Instagram上的使用者。所提出的研究框架包括三個階段，首先，我們從Instagram蒐集資料，並提取照片、關注和社群相關的特徵；接著，我們從IMDb蒐集資料，並提取文字、海報、預告片和社群相關的特徵；最後，根據所提取的特徵，我們提出一個電影帳戶推薦模型（MARM）來計算每個電影的推薦分數，並推薦得分最高的前k部電影給使用者。實驗結果顯示，我們提出的方法優於所有的比較方法，並可減輕冷啟動問題的效應。我們所提出的研究框架可以幫助電影公司或企業吸引潛在觀眾，並擬訂有效的目標式宣傳策略。
並列摘要
With the growth of social networks, movie companies start to create accounts to promote their movies on various social platforms, especially on popular ones like Instagram. Therefore, in this study, we propose a framework to recommend movie accounts to users on Instagram. The proposed framework contains three phases. First, we extract the photo, following and social feature vectors from the data collected from Instagram. Next, we extract the textual, poster, trailer and social feature vectors from the data collected from IMDb. Finally, based on the feature vectors extracted, we propose a Movie Accounts Recommendation Model (MARM) to compute the recommendation score of each movie account, and recommend top-k movies with the highest scores to users. The experimental results show that our proposed method outperforms the state-of-the-art methods in terms of precision, recall, F1-score and Normalized Discounted Cumulative Gain (NDCG), and mitigates the effect of cold start problems. Our proposed framework can help movie companies or businesses reach potential audiences and implement effective targeted advertising strategies.