Polytechnic University of Valencia Congress, CARMA 2020 - 3rd International Conference on Advanced Research Methods and Analytics

Font Size: 
Identification of online reviews helpfulness using Neural Networks
María Olmedilla, Mª del Rocío Martínez Torres, Sergio L. Toral

Last modified: 10-07-2020

Abstract


During the last decade, research has shown that identifying helpful reviews
from a big amount of user-generated review data has been a trending topic.
This study proposes a classification system using an adaptive implementation
of 1D Convolutional Neural Networks (CNNs) that can early identify whether
an online review is helpful, fair or not helpful with 80% of accuracy. After
using the neuronal encoding, a cluster analysis of the helpful and not helpful
was made. The results reveal that the most significant words and documents
for helpful reviews clusters describe cars and their characteristics. Whereas
not helpful reviews clusters express details on car-related shops/companies in
general.


Full Text: PDF