Multiple aspect ranking for opinion analysis

Abstract

We address the problem of analyzing multiple related opinions in a text. For instance, in a restaurant review such opinions may include food, ambiance and service. We formulate this task as a multiple aspect ranking problem, where the goal is to produce a set of numerical scores, one for each aspect. We present an algorithm that jointly learns ranking models for individual aspects by modeling the dependencies between assigned ranks. This algorithm guides the prediction of individual rankers by analyzing meta-relations between opinions, such as agreement and contrast. We provide an online training algorithm for our joint model which trains the individual rankers to operate in our framework. We prove that our agreement-based joint model is more expressive than individual ranking models, yet our training algorithm preserves the convergence guarantees of perceptron rankers. Our empirical results further confirm the strength of the model: the algorithm provides significant improvement over both individual rankers, a state-of-the-art joint ranking model, and ad-hoc methods for incorporating agreement.