Abstract
Document recommendation involves the recommendation of documents similar to those that a user has preferred in the past. The Vector Space Model (VSM) is commonly adopted to denote the document objects and user interests. The user interests are extracted from the documents that a user has browsed. The interest degree of the user is calculated using the TF-IDF method, but the time factor is not considered. The recent documents that a user has browsed embody much more his/her interests. This study proposes a time-aware and grey incidence theory based user interest model to improve document recommendation. First, the time-aware user interest model is proposed based on the analysis of the user interests, document objects and user interest knowledge table. Second, a coefficient vector model of the user interest degree is designed using the grey incidence theory to differentiate the main from the minor user interests. The time-aware and grey incidence theory based user interest model is then exploited to produce document recommendations. Finally, the experiment and evaluation metrics are studied. The results show that the model proposed outperforms other related models and recommends more accurate documents to the users.