With the rise of mobile terminals and the maturity of positioning technology, the amount of available spatiotemporal data continues to grow rapidly, so it is crucial to be able to process it efficiently. This paper proposes a multi-level indexing technique based on context dimension awareness. It first selects the partition order by the unit scale of each context dimension in the dataset. Second, it consider the distribution of each context dimension in the dataset, choose an appropriate partitioning method, and divide the dataset into multiple balanced splits. We test the method on real-world datasets, and experiments show that the speed of query execution increased and resource-use efficiency improved by our approach.<br>

With the rise of mobile terminals and the maturity of positioning technology, the amount of available spatiotemporal data continues to grow rapidly, so it is crucial to be able to process it efficiently. This paper proposes a multi-level indexing technique based on context dimension awareness. It first selects the partition order by the unit scale of each context dimension in the dataset. Second, it consider the distribution of each context dimension in the dataset, choose an appropriate partitioning method, and divide the dataset into multiple balanced splits. We test the method on real-world datasets, and experiments show that the speed of query execution increased and resource-use efficiency improved by our approach.<br>