Özet:

          With rapid advance of data mining technologies, large scale data processing is largely considered to be improving the quality of e commerce industries.  Despite of popularity, data processing and management has become primary importance to handle the streaming data in e commerce applications. It is one of the large scale and potential application growing around the world. In order to provide an efficient solution to handle large scale distributed application data, Top K Multikeyword search provision has been modelled to provide fine-grained keyword data search over the large scale distributed data at the particular time using searchable models. Nevertheless, Top K Multikeyword scheme already developed is unable to support fine-grained data search in an effective manner and fails to provide accurate data retrieval over streaming data. In order to manage those issues, in this paper, a new fine-grained ranked multi-keyword distributed data search scheme over streaming data has been proposed by developing a data intensive –distributed processing paradigm using deep learning architectures has been proposed for along leveraging efficient ranked multi-keyword search methodologies. In Specific, popular data representation model named as vector space model and the term representation model named as TF*IDF model are combined to form the data index construction, data query vector construction and automatic trap door generation. In this distributed multi-keyword top- k data search mechanism has been constructed in order to improve the data query efficiency for data retrieval by simultaneously supporting dynamic update operations on the distributed servers. A deep learning algorithm has been applied to process the data records, index, query vector. Meanwhile the proposed model ensures accurate data relevance score computation between data index and data query vectors. Finally it enables users to gain a broad scope of access to their retrieved data based on their criteria’s, in that high level criteria never requires dynamic strategies for retrieving from the data servers. The proposed scheme for data retrieval has been validated extensively in terms of accuracy and efficiency and its performance of the model has found to be good and excellent model in comparison against other state of art schemes and it offers high level of data retrieval with shortened execution time.

Anahtar Kelimeler: