Performance comparison of case retrieval between Case Based Reasoning and Neural Networks in Predictive Prefetching

摘要

Cache being a fastest medium in memory hierarchy has a vital role to play in memory hierarchy but cannot comprehend speed disparity of processor and memory alone. Predictive prefetching being one of the major concerns in computing systems. The higher level of predictive accuracy is greatly desired. In order to improve the predictability we are looking forward to benefit hybrid of case based reasoning and neural networks. But the most important aspect in this hybrid approach is that of case retrieval which yields related solutions to current problem. We have shown and proved that neural networks have better predictive performance than CBR while performing case retrieval.