Abstract: Sub-prevalent co-location pattern is a subset of co-location patterns, and its mining is based on the star instances model.
In the absence of feature utility values and instance utility values in instance sets, this study investigates how to calculate the utility
value of star participating instances based on the star instances model, and proposes a method for calculating the feature star utility
participation rate. In co-location pattern dominant feature mining, existing research rarely considers the influence of distance
distribution between feature instances on dominant relationships. Therefore, this study explores how to use the distance distribution
between star participating instances and neighbor instances as the dominant weight between features, and proposes a weighted method
for calculating the feature star dominance rate. Finally, an algorithm for mining spatial high utility sub-prevalent co-location patterns
with dominant features is proposed. The experimental results show that this algorithm can effectively distinguish the utility values and
dominant features of patterns, and its effectiveness is significant.

Key words: co-location patterns, sub-prevalent co-location pattern, feature star utility participation rate, weighted feature dominance
rate, utility value of star participating instances