A Recurrent Neural Network for Sentiment Quantification

TL;DR

This paper introduces QuaNet, a recurrent neural network designed for sentiment quantification that leverages classification predictions to improve accuracy, outperforming existing methods.

Contribution

The paper presents QuaNet, a novel RNN architecture that learns higher-order quantification embeddings and refines them using simple classify-and-count predictions.

Findings

QuaNet significantly outperforms state-of-the-art baselines in sentiment quantification.

The approach effectively combines classification predictions to improve quantification accuracy.

Experimental results demonstrate the superiority of QuaNet on real-world sentiment datasets.

Abstract

Quantification is a supervised learning task that consists in predicting, given a set of classes C and a set D of unlabelled items, the prevalence (or relative frequency) p(c|D) of each class c in C. Quantification can in principle be solved by classifying all the unlabelled items and counting how many of them have been attributed to each class. However, this "classify and count" approach has been shown to yield suboptimal quantification accuracy; this has established quantification as a task of its own, and given rise to a number of methods specifically devised for it. We propose a recurrent neural network architecture for quantification (that we call QuaNet) that observes the classification predictions to learn higher-order "quantification embeddings", which are then refined by incorporating quantification predictions of simple classify-and-count-like methods. We test {QuaNet on sentiment quantification on text, showing that it substantially outperforms several state-of-the-art baselines.