Probabilistic reverse top-k (RTk) queries are an important tool for market analysis, but existing algorithms are limited by computationally intensive preprocessing. This paper introduces Probabilistic Partition-based Processing for Reverse queries (PPPR), a framework that addresses this bottleneck. PPPR combines density-based user clustering with a cluster-aware sampling strategy to efficiently estimate user-specific thresholds, replacing exact computation with a data-aware approximation. These estimated thresholds enable a hierarchical pruning strategy that uses cluster centroids for coarse-grained filtering, followed by user-level refinement. Experimental evaluation on large-scale synthetic datasets (50K–150K users) and real-world hotel booking data demonstrates that PPPR reduces preprocessing and query execution times compared to state-of-the-art baselines, achieving 94–97% average accuracy with 85% of queries maintaining >90% accuracy. While accuracy exhibits variance for rare items (<5% of queries), this efficiency-fidelity trade-off makes PPPR suitable for applications prioritizing speed and scalability.