Combination of Feature-based and Instance-based methods for Domain Adaptation in Sentiment Classification

Methods of domain adaptation can be roughly divided into two categories: feature-based and instance-based. In summary, both methods are a kind of weighted-learning, but feature-based gives weight to features and instance-based gives weight to instance. Generally, feature-based is more effective than instance-based. However, these two methods can be combined to improve the accuracy of only the feature-based method. In this paper, we do it using the neural network model, where we use the feature-based method as SVD and instance-based method proposed in our previous work. In the experiment for the Amazon dataset, we confirmed the effectiveness of the proposed method.