Attribute Sentiment Scoring with Online Text Reviews: Accounting for Language Structure and Attribute Self-Selection

Abstract

The authors address two novel and significant challenges in using online text reviews to obtain attribute level ratings. First, they introduce the problem of inferring attribute level sentiment from text data to the marketing literature and develop a deep learning model to address it. While extant bag of words based topic models are fairly good at attribute discovery based on frequency of word or phrase occurrences, associating sentiments to attributes requires exploiting the spatial and sequential structure of language. Second, they illustrate how to correct for attribute selfselection-reviewers choose the subset of attributes to write about-in metrics of attribute level restaurant performance. Using Yelp.com reviews for empirical illustration, they find that a hybrid deep learning (CNN-LSTM) model, where CNN and LSTM exploit the spatial and sequential structure of language respectively provide the best performance in accuracy, training speed and training data size requirements. The model does particularly well on the "hard" sentiment classification problems. Further, accounting for attribute self-selection significantly impacts sentiment scores, especially on attributes that are frequently missing.

A Comparison and Interpretation of Machine Learning Algorithm for the Prediction of Online Purchase Conversion

“…For example, Liu et al [48] analyzed consumer reviews and reported that aesthetics and price influence conversion. In addition, Chakraborty et al [49] developed a Hybrid CNN-LSTM model to extract emotional characteristics from text data, and showed that it solves difficult emotion classification problems well for Yelp reviews. Zhang et al [50] analyzed the effect of images on Airbnb's accommodation demand by using deep learning.…”

Text vs. Image: An application of unsupervised multi-modal machine learning to online reviews

“…The surge of unstructured data sources has made methods for translating textual and visual content into useful information increasingly necessary. Online reviews have shown to be an incredibly rich source of unstructured information about consumer perceptions and experiences that provide a peek into the minds of its producers and receivers [3,4,5,6,7]. In this research, we sought to improve our understanding of the different modalities that constitute an online review -text and imagery.…”

“…It allows users to share their experiences with comments, reviews, pictures, ratings on a hotel, destination, or any tourist attraction. These reviews have been shown to be a valuable source of information [3,4,5,6,7] Future consumers resolve uncertainty about the quality of a hotel and its facilities by reading the online reviews given by customers on a website [1]. In addition, expectations of customers usually do not conform with their experience of a hotel because of varied reasons [2].…”

“…A recent stream of research has focused on investigating the textual content in online user-generated content such as online reviews. These studies range from opinion summarization [9,10], sentiment analysis [11], extracting business-specific information [1,12], or a combination such as fine-grained attribute sentiment analysis [7]. These methods are helpful in easing online decision-making by summarizing the most useful information to consumers [9].…”

“…In addition, the methods are necessary for firms to recover firm-quality information from reviews [12], or even infer quality information from what consumer do not discuss in their reviews [7].…”