To help customers with their buying decisions, many e-commerce websites allow buyers to provide reviews for their purchased products; however, due to a large amount of reviews for many similar online products, consumers often feel it is difficult to determine which products have the most desirable features. In this paper, we propose a supervised learning approach to efficiently and effectively analyzing online product reviews and identifying the strengths and weaknesses of a product by its product features. The proposed approach uses a novel Feature-based Sentence Model (FSM), where a latent layer, called the feature layer, is introduced between review sentences and words. Once a model has been trained with sufficient labeled data points, it can identify the most related product feature, if any, for each review sentence. With the identified product features, we perform sentiment analysis for each sentence, and derive the weighted feature preference vectors for the review. Finally, we combine the results of all review comments for a product into a review summary. We demonstrate in two case studies that our approach works more effectively than existing approaches, and provides consumers a much easier way to find online products with the most desirable product features.