A novel activity recommendation for tourists based on location using recurrent neural network in comparison with convolutional neural network with improved accuracy Available to Purchase

The goal is to improve the innovative Tourist Activity Recommendation System by utilizing Recurrent Neural Networks algorithm, as opposed to the more traditional Conventional Neural Networks algorithm, considering the user’s present location. Materials and Methods: In order to forecast the identification of extremist reviews, this study employs Convolutional Neural Networks (CNNs) and Recurrent Neural Networks (RNNs) with different grounding categories. The kaggle website data csv file repositories are used to compile the dataset presented in this article. With α = 0.05 and power = 0.80 as the parameters, a G-Power test was carried out, leading to a power level of 80%. With a total of 5,456 samples, Recurrent Neural Networks (RNNs) and Convolutional Neural Networks (CNNs) improve the accuracy rate of predicting extremist reviewer groups in online commerce. Results: The results showed that the new activity recommendation system had an accuracy of 94.92% with new Recurrent Neural Networks and an accuracy of 93.26% with Convolutional Neural Networks. The significance value for the independent sample-t-test was p=0.042 (p<0.05). Between two algorithms, there is a statistically significant relationship. Therefore, the Recurrent Neural Network algorithm, rather than the CNN algorithm, will carry out the task of providing tourists with innovative activity recommendations based on their current location.