Abstract:<p>Personalized Point-of-Interest (POI) Recommendation is very important for application platforms based on Location Based Social Networks (LBSNs). It can assist users in making decisions to alleviate the problem of information overload, and can also improve the user experience of these platforms and advance platform operators achieve personalized and accurate advertising. However, there exist some problems of data sparseness and cold start for a single user, and it is also difficult to mine valuable long-tailed POIs, although the size of the check-in data is large. Therefore, in order to address the above problems, we propose a personalized POI Recommendation approach based on Hierarchical Attention Mechanism (HAM-POIRec) which can effectively increase data utilization. Firstly, we define the concepts of explicit features and implicit features, which pave the ideas of selecting data and computational models for POI recommendation based on machine learning. Secondly, we propose a hierarchical attention mechanism with the structure of local-to-global, which extracts contributions and mines more hidden information from individual features, combination features, and overall features. Finally, we present the Natural Language Processing (NLP)-based "User-POI" matching mechanism for the first time in the field of POI recommendation to improve the recommendation accuracy by fine-tuning the POIs predicted by the recommendation system. Extensive experiments are conducted for demonstrating that the HAM-POIRec method outperforms state-of-the-art DeepPIM method and the other comparison methods (SAE-NAD, MGMPFM and LRT), especially in predicting sequence POIs and solving cold start problem.</p>

What problem does this paper attempt to address?

The problem that this paper attempts to solve is how to perform Point - of - Interest (POI) recommendation more efficiently in Location - Based Social Networks (LBSNs). Specifically, the paper focuses on the following challenges: 1. **Data Sparsity**: Since users' check - in data is usually very sparse, it is difficult to accurately capture users' preferences. 2. **Cold - Start Problem**: For new users or new POIs, due to the lack of sufficient historical data, it is difficult for the recommendation system to provide accurate recommendations. 3. **Long - Tail POI Mining**: Although the amount of check - in data is large, how to effectively mine valuable long - tail POIs (i.e., uncommon POIs) remains a challenge. 4. **Time - Series Features**: Users' interests change over time. For example, restaurants in the morning may be different from those in the evening. 5. **Social Relationship Influence**: Users' choices are influenced by their social circles. For example, friends' preferences will influence users' decisions. 6. **Utilization of Text and Image Information**: How to extract useful information from users' comments and POI descriptions to improve the accuracy of recommendations. To address the above challenges, the paper proposes a personalized POI recommendation method based on the hierarchical attention mechanism (HAM - POIRec). The main contributions of this method include: 1. **Definition of Explicit and Implicit Features**: Explicit features are extracted from structured data, such as user ID, time, location, etc.; implicit features are extracted from unstructured data, such as user comments and POI descriptions. 2. **Hierarchical Attention Mechanism**: Through the local - to - global hierarchical attention mechanism, more hidden information is extracted from individual features, combined features, and overall features, thereby improving the utilization rate of sparse data. 3. **NLP - based "User - POI" Matching Mechanism**: For the first time, natural language processing (NLP) technology is introduced in the POI recommendation field. The semantic similarity between user comments and POI descriptions is calculated to improve the accuracy of recommendations. Through these innovations, the paper aims to improve the accuracy of POI recommendations and user experience, especially in predicting sequential POIs and solving the cold - start problem.

Efficient point-of-interest recommendation with hierarchical attention mechanism

Influence-Aware Successive Point-of-Interest Recommendation

GSSM: An Integration Model of Heterogeneous Factors for Point-of-Interest Recommendation

A Temporal-Aware POI Recommendation System Using Context-Aware Tensor Decomposition and Weighted HITS.

User Multi-behavior Enhanced POI Recommendation with Efficient and Informative Negative Sampling.

A Multi-Task Embedding Based Personalized POI Recommendation Method

A BiLSTM-attention-based point-of-interest recommendation algorithm

CHA: Categorical Hierarchy-based Attention for Next POI Recommendation

Hierarchical and Multi-resolution Preference Modeling for Next POI Recommendation

A Context Awareness Hierarchical Attention Network for Next POI Recommendation in IoT Environment

Exploring Temporal and Spatial Features for Next POI Recommendation in LBSNs

Sequential-hierarchical attention network: Exploring the hierarchical intention feature in POI recommendation

P3OI-MELSH: Privacy Protection Target Point of Interest Recommendation Algorithm Based on Multi-Exploring Locality Sensitive Hashing

POI recommendation method using LSTM-attention in LBSN considering privacy protection

Intent-aware Graph Neural Network for Point-of-Interest Embedding and Recommendation

ST-PIL: Spatial-Temporal Periodic Interest Learning for Next Point-of-Interest Recommendation

STP-UDGAT: Spatial-Temporal-Preference User Dimensional Graph Attention Network for Next POI Recommendation

Next POI Recommendation Based on Location Interest Mining with Recurrent Neural Networks

Exploiting Category-Level Multiple Characteristics for POI Recommendation

A Next POI Recommendation Based on Graph Convolutional Network by Adaptive Time Patterns

Empowering Next POI Recommendation with Multi-Relational Modeling