FINDING NOVELTY IN ONLINE TRAVEL REVIEWS USING DEEP LEARNING

Abstract

The exploration intends to make a classification structure and DL model, BERT-BiGRU, that can automatically detect novelty seeking (NS) from online travel reviews. These assessments are enormous and unstructured, making manual arrangement troublesome. Programmed approaches to productively deal with and assess the gigantic material are required because of this limitation. The proposed DL model, BERT-BiGRU, based on Bidirectional Encoder Representations from Transformers, has accomplished great accuracy and F1 scores while identifying the NS character quality from surveys. This shows the model's viability. The analysis demonstrates the way that strong computational strategies can consequently recognize character attributes, for example, curiosity chasing, from trip assessments. Mechanization is more proficient and versatile than human methodologies. The undertaking's discoveries give a total NS character trademark classification framework. This worldview might be utilized in the travel industry advertising and proposal frameworks to grasp client inclinations. The drive propels brain science and showcasing computational strategies. The utilization of DL models like BERT-BiGRU to distinguish character attributes from unstructured text information advances progressed approaches in numerous disciplines, showing their commitment past customary applications. Project developments incorporate a hybrid model called "BERT CNN-BI-GRU" and "BERT-LSTM-GRU" with close to 100%