Abstract:Conventional prefetching methods that look at access correlations operate under the presumption that two files that were previously viewed often and in close proximity will probably be accessed together shortly. Access correlation detection techniques primarily use frequent sequence mining to find recurrent access patterns. Since establishing patterns of access might have a substantial impact on the accuracy of future access predictions, it might be exceedingly challenging. Thus, it is important to choose values for confidence, gap, and lowest level support carefully in order to exclude low-quality access correlations. This work presents a novel DFS model using dynamic prefetching, which consists of four processes: (i) popular file identification (ii) improved support value estimation, (iii) extraction of frequent block access patterns (FBAP), and (iv) matching algorithm, Essentially, a new threshold computation is used to determine the most common input file sizes. New threshold-based FBAPs are extracted in the next phase after the support value has been calculated. Lastly, weight-optimized Convolutional Neural Network (CNN) and Long Short Term Memory (LSTM) based on Upgraded Shark Smell with Opposition Solution (USS-OS) are employed for matching. For the best case scenario, the proposed method achieves higher value of (~ 0.98) as compared to the existing models such as HC + Arithmetic Optimization Algorithm (AOA) (~ 0.96), HC + Bald Eagle Search (BES) (~ 0.97), HC + Shark Smell Algorithm (SSA) (0.97), HC + Hunger Games Search (HGS) (0.97), and HC + Harris Hawks Optimization (HHO) (~ 0.95) respectively. Finally, the studies are conducted to analyze the developed model.

Mining Infrequently-Accessed File Correlations In Distributed File System

Mining Associated and Item-Item Correlated Frequent Patterns

A Load Balancing Strategy Based on Data Correlation in Cloud Computing

A Distributed Method for Fast Mining Frequent Patterns From Big Data

File fetching in distributed file system via optimization assisted hybrid deep learning model

LogMaster: Mining Event Correlations in Logs of Large scale Cluster Systems

Distributed Storage Algorithm for Geospatial Image Data Based on Data Access Patterns

A STABLE PARALLEL DISTRIBUTED FREQUENT ITEMSET MINING ALGORITHM AND ITS APPLICATION

Distance based prefetching algorithms for mining of the sporadic requests associations

Mining Both Associated and Correlated Patterns

Sequential Data Mining using Correlation Matrix Memory

Parallel Frequent Pattern Discovery: Challenges and Methodology

Efficiently mining mutually and positively correlated patterns

Reducing False Discoveries in Statistically-Significant Regional-Colocation Mining: A Summary of Results

Tree Partition Based Parallel Frequent Pattern Mining on Shared Memory Systems

Correlated Utility-based Pattern Mining

MobileCopy: Improving Data Availability and File Search Efficiency in Delay Tolerant Networks against Correlated Node Failure

Data Sharing Analysis Of Emerging Parallel Media Mining Workloads

CODEM: A Novel Spatial Co-location and De-location Patterns Mining Algorithm

A Connectivity Reduction Based Optimization Method for Task Level Parallel File System

Tight Correlated Item Sets And Their Efficient Discovery