Hot data prediction method based on joint optimization echo state network

Abstract

The invention discloses a hot data prediction method based on a joint optimization echo state network. Quantum particle swarm optimization is creatively used for calculating storage layer parameters of an echo state network to obtain optimal storage layer parameters; in a particle position updating process, an output weight is calculated by combining an echo state network of L2 + self-adaptive L1 / 2 regularization constraint and a global optimal adaptive value is calculated; at the end of iteration, the particle position corresponding to the global optimal adaptive value is taken as an optimalstorage layer parameter; and finally, based on the optimal storage layer parameter, a final output weight is calculated by adopting an echo state network of L2 + adaptive L1 / 2 regularization constraint. The hot data is predicted by using the final output weight and the logic block address where the input historical hot data is located. According to the method, hot data prediction replaces hot dataidentification, and a predictive hot data identification technology is disclosed and can better serve garbage collection and wear leveling tasks of the solid state disk.

Description

technical field [0001] The invention belongs to the technical field of chaotic time series prediction, and in particular relates to a thermal data prediction method based on a jointly optimized echo state network. Background technique [0002] As a non-volatile storage technology, NAND flash memory is widely used in communication systems and consumer electronics, and it has higher access speed and power efficiency than hard disk drives. In NAND flash-based consumer electronics devices, a large number of applications rely on NAND flash for data exchange, file storage, and video storage. NAND flash memory is mainly used to store large-capacity data. The NAND structure can provide extremely high cell density, which can achieve high storage density, high writing and high erasing speed; therefore, NAND flash memory is mostly used for large-capacity data storage, such as solid-state drives. In the future, the demand for NAND flash memory will continue to grow, and it will be main...

AI Technical Summary

Problems solved by technology

Another serious problem is that the accuracy of identification is not high. The hot data identification mechanism based on Bloom filter is widely applicable to the hot and cold data identification of SSD, but the inherent defect of Bloom filter is false positive, that is, the data that does not belong to the set Incorrectly determined to be in the set

In addition, the hot data identification method based on load request size and load access mode considers relatively single considerations, fails to fully consider the local characteristics of the load, and the accuracy of hot data identification is not high

Images