在WEB2.0的時代，網際網路中充斥著各式各樣的互動式平台。就音樂網站而言，使用者除了聽音樂外，更開始習慣於虛擬空間中交流及分享意見，並且在這些交流、分享的過程中留下他們的足跡，間接的提供許多帶有個人色彩的資訊。利用這些資訊，更貼近使用者的推薦系統因應而生。本研究中，將針對使用者過去存取過的音樂特徵以及使用者於系統中留下的文字評論特徵這兩個部份的資料，做音樂特徵的擷取、找尋具有價值的音樂特徵區間、建立使用者音樂特徵偏好，以及文字特徵的擷取、建立使用者文字特徵偏好。接著，採用協同式推薦方式，將具有相同興趣的使用者分於同一群，推薦給使用者與之同群的使用者的喜好物件，但這些推薦之物件為該使用者過去並沒有任何記錄於這些喜好物件上之物件。我們希望對於音樂推薦考慮的開始不只是音樂上之特徵，更包含了使用者交流、互動中留下的訊息。 / In the era of Web2.0, it is flooded with a variety of interactive platforms on the internet. In terms of music web site, in addition to listening to music, users got used to exchanging their comments and sharing their experiences through virtual platforms. And through the process of exchanging and sharing, they left their footprints. These footprints indirectly provide more information about users that contains personal characteristics. Moreover, from this information, we can construct a music recommendation system, which provides personalized service. In this research, we will focus on user’s access histories and comments of users to recommend music. Moreover, the user’s access histories are analyzed to derive the music features, then to find the valuable range of music features, and construct music profiles of user interests. On the other hand, the comments of users are analyzed to derive the textual features, then to calculate the importance of textual features, and finally to construct textual profiles of user interests. The music profile and the textual profile are behaviors for user grouping. The collaborative recommendation methods are proposed based on the favorite degrees of the users to the user groups they belong to.

隨著手持裝置的普及與社群網路的興起，大眾可以隨時拍攝影片並且上傳至網路上與他人分享。但是一般使用者產生的影片若少了配樂，將失色許多。除了原本影片帶給人們的視覺觀感之外，配樂可以帶給人們聽覺的觀感，進而使得人們可以更容易了解影片的情感，也可以讓人們更能夠融入在影片中。背景音樂推薦的研究主要有兩大種做法，Emotion-mediated Approach與Correlation-based Approach。我們使用Correlational-based Approach的方法，利用Correlation Modeling找出影片特徵值與音樂特徵值之間的關係。但是由於目前Correlation-based Approach的研究只有考慮到全域特徵，因此在此論文中，我們提出了區域特徵。區域特徵利用時間序列表達影片細部的變化，並且將區域特徵與全域特徵結合至Correlation Modeling中，透過 MLSA、CFA、CCA、KCCA、DCCA、PLS、PLSR演算法找出其中的關係並且產生背景音樂推薦的Ranking List，實驗部份比較了各個演算法在背景音樂推薦上的準確率，並且觀察Global Features與Local Features之間的準確率。 / Background music plays an important role in making user-generated video more colorful and attractive. One of current research on automatic background music recommendation is the correlation-based approach in which the correlation model between visual and music features is discovered from training data and is utilized to recommend background music for query video. Because the existing correlation-based approaches consider global features only, in this work we proposed to integrate the temporal sequence of local features along with global features into the correlation modeling process. The local features are derived from segmented audiovisual clips and can represent the local variation of features. Then the temporal sequence of local features is transformed and incorporated into correlation modeling process. Cross-Modal Factor Analysis along with Multiple-type Latent Semantic Analysis, Canonical Correlation Analysis, Kernel Canonical Correlation Analysis, Deep Canonical Correlation Analysis, Partial Least Square and Partial Least Square Regression, are investigated for correlation modeling which recommends background music in ranking order. In the experiments, we first compare the results of only global features, only local Features and incorporating global and local Features among each algorithm. Then second compare the results of different clip numbers and Fourier coefficients.