Similarity-Based Clustering and Security Assurance Model for Big Data Processing in Cloud Environment
Parthiban Krishnamoorthy, SUJATHA SUNDARAM · ECONOMIC COMPUTATION AND ECONOMIC CYBERNETICS STUDIES AND RESEARCH · 2018
Present internet users evidently register the adoption of services offered through web resources.Presently prevailing scenario of adopting a cloud environment assists this practice in an intensified manner.Consequently, this influences the fabrication of an expertise web service recommendation system that serves with an ability to reach an utmost proficiency in retrieving web services.A prevailing search engine usually opts for providing recommendations for the query being posed by the user through utilization of a prevalent Collaborative Filtering methodology.In case of a accomplishing a web service from a cloud paradigm, the intricacy arises in processing the vast amount of information within a stipulated amount of time.However, the utmost challenge lies with the resource provisioning for those tasks demanded by user.In addition, effectual performance of the entire system gets degraded out of furnishing request in view of precedent user rating alone in an implicit way.In order to surmount over the issue of accomplishing an unrelated web service from a search engine, an inventive approach devised as Distance based Agglomerative Clustering and Secure Service Recommendation (DAC-SSR).A robust form of subset formulation via user-item and item-item similarity promotes the existence of resourceful associative similarity in between that user query and web service.Preprocessed dataset fed to the consecutive models comprises of various domains and proceed with the distance based agglomerative clustering.Those clusters get subjected to adopt with Similarity-ranking (S-r) scheming that tends to rank those clusters relying upon the similarity assessed.However, a robust user-item similarity also gets computed through the implication of a cosine similarity.Afterward, these ranked datasets acquire its position in a server as encrypted data by means of realizing Elgamal cryptosystem.Realizing efficacy of DAC-SSR gets accomplished through a comparative analysis of an existing methodology OGRPL-FW in terms of RMSE, MAE and running time metrics.