Incorporating a Topic Model into a Hypergraph Neural Network for Searching-Scenario Oriented Recommendations

Abstract

The personalized recommendation system is a useful tool adopted by e-retailers to help consumers to find items in line with their preferences. Existing methods focus on learning user preferences from a user-item matrix or online reviews after purchasing, and they ignore the interactive features in the process of users’ learning about product information through search queries before they make a purchase. To this end, this study develops a topic augmented hypergraph neural network framework to predict the user’s purchase intention by connecting the latent topics embedded in a consumer’s online queries to their click, purchase, and online review behavior, which aims at mining the connection information existing in the interaction graph domain. Meanwhile, in order to reduce the influence of text noise words by fusing topic information, we integrate the topic distribution and convolutional embedding to better represent each user and item, which can make up for the lack of topic information in traditional convolutional neural networks. Extensive empirical evaluations on real-world datasets demonstrate that the proposed framework improves the novelty of recommendation items as well as accuracy. From a managerial perspective, recommending diversified and novel items to consumers may increase the users’ satisfaction, which is conducive to the sustainable development of e-commerce enterprises.

Advancing Complexity Research in Earth Sciences and Geography

“…In their study, Tao et al [16] develop a hierarchical maximal accessibility equality model to examine the equality of accessibility to healthcare services in Shenzhen, China. In addition, Huang and Liu [17] propose a more accurate personalized recommendation system for e-retailers that is also computationally more efficient. While all this research is fascinating, it would be desirable to see whether the results of these studies can be applied in practice and make a profound impact on society.…”

“…This book arises from a Special Issue of Applied Sciences that aimed to systematically examine the many complex phenomena that occur in earth sciences and geography, employing state-of-the-art methods for modeling complex data in order to invigorate research in earth sciences and geography, and to facilitate the further development of complexity science. Altogether, this Special Issue comprises 20 papers, contributed by researchers from all over the world and covering a range of diverse topics, including the encryption of digital elevation models [2], facies heterogeneity [3], the simulation of the snow cover process [4], the exploration of ice elevation change [5], earthquake and seismic activity [6][7][8][9], landslide susceptibility [10,11], the effect of reforestation [12], coordination between the supply and demand of ecosystem services [13], indoor positioning [14], public transport flow networks and retail store locations [15], the equality of healthcare facilities [16], recommender systems for e-retail [17], globalization [18], international trade and optimal industrial structure [19], risk analysis [20], and the quantification of political processes [21]. Below, I briefly explain the premise of each work, and when appropriate, highlight what could be further explored in future.…”

A Contrastive News Recommendation Framework based on Curriculum Learning

“…Recently, several researchers have successfully incorporated deep learning techniques within the news recommendation. During this process, the deep learning model learns and extracts significant features from the source data to offer tailored recommendations to users according to their actions and preferences [5,6]. For example, Recurrent Neural Networks and Transformers, recognized for their efficacy in sequence modeling, have been employed to accurately model user interests in various studies [7][8][9].…”

Deep Learning-Based Recommendation System: Systematic Review and Classification

“…Deep learning models can learn and extract relevant features from raw data, making them highly effective in handling complex and highdimensional datasets [34]. This capability is particularly relevant in recommendation systems, where personalized recommendations can be made to users based on their behavior and preferences [35], [36].…”