| Summary: | 碩士 === 中原大學 === 資訊管理研究所 === 107 === The amount of information on the Internet is huge, so that people have difficulty to deal with. Users also need to spend a lot of time to find the needed information. Information overload problem becomes a significant issue for users and online businesses. To resolve this problem, recommender systems are proposed in many researches or applications. The widely used method in recommender systems is the user-based collaborative filtering. It only concerns users’ rating history to analyze users’ preferences. However, users’ text data may contain in users’ preference or sentiment information. Such information is used in analyzing user’s preference more precisely. In this work, we proposed a method which is called Aspect-based Deep Learning Rating Prediction Method (ADLRP). This method can extract the aspects, sentiment and semantic information from users’ and items’ reviews. Then, the deep learning method is used to generate users’ and items’ latent factors. According to these three features, the matrix factorization method is applied to make rating predictions for items. The experimental results show that the proposed method performs better than the traditional methods of rating prediction and conventional artificial neural networks. The proposed methods can precisely and efficiently extract the sentiment and semantic of each aspect from review texts, and enhance the prediction performance of rating predictions.
|
|---|
