.png){kind=icon}
.png){kind=icon}
Return
Multisource Graphs and Dual KAN-Transformers for Next POI Recommendation
.png){kind=icon}
PRE
AI
J
Z
王
Q
L
X
DOI：10.1109/jiot.2026.3704841.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Next point-of-interest (POI) recommendation aims to predict a user’s subsequent location based on their check-in sequence, thereby supporting real-time decision-making for mobility and local services. However, real-world scenarios are constrained by spatial, temporal, and contextual factors, which complicate unified modeling of user preferences and degrade predictive reliability. Additionally, these scenarios also face problems such as sparse user check-in data and nonstationary preferences. Therefore, we propose a framework based on multisource graphs and dual KAN-Transformers (MSG-DKT) for next POI recommendation. Concretely, MSG-DKT is composed of three modules: 1) <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">short-term preference learning:</i> Based on multisource graphs and contextual information, a short-term preference is constructed. The S-KAN-Transformer (SKT) then jointly predicts the next POI, its category, and the visit time; 2) <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">long-term preference learning:</i> The long-term sequence is enhanced by the L-KAN-Transformer (LKT) and aggregated by long-term self-attention into a stable preference representation; and 3) <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">hierarchical fusion:</i> The outputs of the short- and long-term branches are fused hierarchically to yield the final prediction. Extensive experiments on two real-world datasets demonstrate the superiority of MSG-DKT over state-of-the-art methods for next POI recommendation.
Keywords:
Dual Kolmogorov–Arnold network (KAN)-Transformers
long-term self-attention
multisource graphs
next point-of-interest (POI) recommendation
Journal
IF:
8.9
Papers:
1.4W
Citations:
7.8W
Organization
Cited Papers