A Novel Redundant Data Retrieval Model based on Parallel Batch Algorithm
Wei Cui, Lu Qingbo, Ding Eng Na, Cui Congli · 2023
For the effective information retrieval models, traditional unsupervised cross-modal hashing methods try to learn hash functions from the underlying structure, distribution, and topological information of the data in order to preserve the original data spatial structure. Hence, to conduct more efficient analysis, this research study proposes a novel redundant data retrieval model based on parallel batch algorithm. Batch parallel proofreading is oriented to large-scale text offline proofreading scenarios, and has powerful computing power provided by the cluster computing. With this operation, the data will be pre-processed to better clean the row data. Then, the hash model is integrated to consider the key as a divisor, the dividend as a maximum prime number p not greater than the length of the hash table m, and uses the remainder of the calculation result as the final hash function. Furthermore, the model is revised to construct the efficient data retrieval task. Through testing, the performance of the proposed algorithm is proven to be efficient.