Enhancing Consumer Sentiment Analysis and Product Improvement with Aspect-Optimized Adaptive GCN (AOA-GCN)

The exponential growth in online reviews offers a company the most precious source of information regarding the customers' emotions and potential product improvement areas. However, it is very challenging to derive meaningful information from the unstructured data. The work presents a new model called Aspect-Optimized Adaptive GCN (AOA-GCN), which combines hierarchical structuring with aspect-based optimization for effective consumer sentiment analysis. The AOA-GCN applies a two-level analysis framework that covers both global sentiments and detailed aspect-specific sentiments in order to build a full understanding of the customer review. The model is essentially based on an Adaptive Lightweight GCN with pruning techniques, ensuring that it does not only reduce computations but also takes into account influential aspects along with its high accuracy, even across extended datasets. Dynamic attention in AOA-GCN will help make critical product features in the customer reviews focus on the issues so improvements can be made on those products specifically. Experimental results show that the improved performance is actually as high as 6.45% over global sentiment classification, 9.12% over aspect-based sentiment analysis, and 7.35<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">%</sup> over cross-domain generalization, and training time is reduced by 35%. The model also exhibits improved interpretability with clearer insights into customer preferences. The proposed AOA-GCN becomes an effective, efficient, and interpretable business tool for optimizing products through insights from consumers.