Few-shot hotel industry site selection prediction method based on meta learning algorithms and transportation...

Area of Science:

• Urban Planning
• Geographic Information Systems
• Artificial Intelligence

Background:

• Optimizing urban spatial structure and tourism services requires rational hotel location selection.
• Artificial intelligence offers a data-driven methodology for this critical decision-making process.

Purpose of the Study:

• To propose a few-shot hotel location prediction method for star-rated hotels in Tianjin using meta-learning and transportation accessibility.
• To enhance the accuracy and efficiency of hotel location selection in urban environments.

Main Methods:

• A meta-learning algorithm was employed to generate initial hotel location predictions.
• Spatial syntax was utilized to construct a transportation accessibility model for secondary screening.
• An appropriateness distribution map was created based on identified demand levels.

Main Results:

• The meta-model achieved 90.45% classification accuracy and 91.90% location fitting degree, an 11% improvement over baseline models.
• Transportation accessibility was a significant factor, contributing 45% to classification information in star-rated hotel distribution.
• Recommended investment areas include Xiaobailou Street, Dawangzhuang Street, and Wudadao Street.

Conclusions:

• The proposed meta-learning approach demonstrates superior performance in few-shot hotel location scenarios.
• The model provides a practical framework for data-driven hotel location decision-making, optimizing urban development and tourism.
• Future hotel investments should strategically target areas with high transportation accessibility and demand.