STELLAR: Spatial-Temporal Latent Ranking for Successive Point-of-Interest Recommendation

Authors: Shenglin Zhao, Tong Zhao, Haiqin Yang, Michael Lyu, Irwin King

AAAI 2016 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Evaluations on two real-world datasets demonstrate that the STELLAR model outperforms state-of-the-art successive POI recommendation model about 20% in Precision@5 and Recall@5.
Researcher Affiliation Academia Shenglin Zhao1,2, Tong Zhao1,2, Haiqin Yang1,2, Michael R. Lyu1,2, Irwin King1,2 1Shenzhen Research Institute The Chinese University of Hong Kong, Shenzhen, China 2Department of Computer Science & Engineering The Chinese University of Hong Kong, Shatin, N.T., Hong Kong {slzhao, tzhao, hqyang, lyu, king}@cse.cuhk.edu.hk
Pseudocode Yes Algorithm 1 gives the detailed procedure to learn the STELLAR model.
Open Source Code No The paper does not provide a direct link or explicit statement about the availability of open-source code for the described methodology.
Open Datasets Yes We use two check-in datasets crawled from real world LBSNs: one is Foursquare data provided in (Gao, Tang, and Liu 2012) and the other is Gowalla data (Zhao, King, and Lyu 2013).
Dataset Splits Yes In order to make our model effective for future check-ins, we split the tuples into two parts, 80% and 20% according to time sequential order.
Hardware Specification No The paper does not provide specific details about the hardware used for running the experiments (e.g., GPU/CPU models, memory).
Software Dependencies No The paper does not list specific software dependencies with version numbers.
Experiment Setup Yes We set latent dimension as 40, and train different models to get their best performances at appropriate parameters. ... The model has best performance when λ = 0.001. ... For the trade-off of performance and computation cost, we suggest to set dimension d = 40.