There is a high demand to summarize and analyze the opinions in online customer reviews. Sentiment analysis is one of the study fields in this area. Mining comparative opinions is an important application of sentiment analysis. It includes identifying the comparative opinions and the aspects that are compared. It also identifies the sentiment classification of the opinion as positive or negative. This helps businesses to make effective decisions in the development and promotion of their products and services, and to better understand their competitors. Different approaches could be used to address this sentiment analysis application, such as Machine Learning. The application is a multi-label classification problem from a machine learning perspective. This paper presents a case study to evaluate three multi-label machine learning classification techniques in addressing the problem. Empirical experiments are conducted on a domain-independent dataset of online customer reviews from Amazon for the evaluation purpose.