ARMLOWA: aspect rating analysis with multi-layer approach

Autor: Devendra K. Tayal, Amita Jain, Swati Sharma, Sayani Ghosal
Rok vydání: 2021
Předmět:
Zdroj: Progress in Artificial Intelligence. 10:505-516
ISSN: 2192-6360
2192-6352
DOI: 10.1007/s13748-021-00252-4
Popis: Aspect-based sentiment rating analysis is a decisive industry application for consumers and service providers. This research needs an efficient tool that reflects various shades of sentiments. The overall rating computation generally considered the average of all aspect ratings and failed to consider greater emphasis on positive and negative aspects. This research presents a novel multi-layer approach with the basic ordered weighted average (OWA) operator named ARMLOWA. Implementation of OWA operator made comprehensive aggregation that tailored user opinion and intention, and this multi-layer approach has shown the model learning advantageously. The model includes aspect segmentation, aspect rating, and overall sentiment rating computation. The model has picked the hospitality domain review dataset with significant aspects, like value, room, location, cleanliness, and service. The model confers 9.18% decrease in root-mean-square error (RMSE) and 29.67%, 4.98% increase for aspect correlation-inside reviews (PAspect) and across reviews (PReview), respectively.
Databáze: OpenAIRE
Externí odkaz: