Big Data Retrieval Using Locality-Sensitive Hashing with Document-Based NoSQL Database

A locality-sensitive hashing (LSH) method in the document-based NoSQL database is proposed for enhancing the ability of arbitrary reads over the previous methodologies. The proposed hash index improves efficiency by reducing the amount of accessing data for search queries by creating buckets based on hyperplanes. The LSH hashes the input data where similar items with high probability maps to the same bucket. They attempt to decrease the volume of candidate data objects matched when reducing the missed nearest neighbors. The data space is divided with randomly chosen hyperplanes to decrease the volume of candidate objects. The values which are nearer to the boundaries (adjacent to the two sides of the hyperplane) are considered. The bucket label's string length is equivalent to the amount of used hyperplanes. The effect of LSH for bucket size balancing and analysis of the non-indexed, hash index, and global-indexed dataset on MongoDB depicts the pre-eminence of the presented hash index.