Lossy Data Storage Using Event-Rarity-Based Smoothing

The patent applies local quality by differentiating compression rates based on the rarity of events. Rare events (with occurrence probability below a threshold) are compressed at a lower rate to preserve their characteristics, while common events are compressed at a higher rate. This is achieved through the smoothness decision unit that determines subject smoothness based on event rarity, enabling selective preservation of important data characteristics while still achieving overall compression.

The patent applies preliminary action by pre-calculating and storing the occurrence probability of events in advance. The probability calculation unit computes how often each event occurs in the dataset before compression begins. During compression, the smoothness decision unit refers to these pre-calculated probabilities to determine appropriate smoothness values, avoiding the need for complex real-time analysis and enabling efficient rarity-based compression control.

The patent applies parameter changes by dynamically adjusting the smoothness parameter based on event rarity. The smoothness decision unit determines subject smoothness values that are specifically tailored to each event's occurrence probability. This parameter adaptation allows the compression process to preserve characteristics of rare events (higher smoothness) while still achieving volume reduction through compression of common events (lower smoothness), optimizing both reliability and storage efficiency.