Please use this identifier to cite or link to this item:
Title: Big Data Analytics For Frequent Text Pattern And Unsupervised Feature Learning
Researcher: Swain, S.K.
Guide(s): Srinivas Prasad, M. Vamsikrishna, and NageshKolagani,
Keywords: Engineering and Technology,Computer Science,Computer Science Interdisciplinary Applications
University: Centurion University of Technology and Management
Completed Date: 2019
Abstract: For global economics and social changes the big data is considered as a new driver. The data collection globally is attaining a tipping point aimed at vital technical transformations that can create opportunities in decision making, health management, education and finance. While the increasing complexities of data for e.g. Data volume, variety, velocity and veracity, the real impact is on our ability to disclose the `valueand#8223; in the data by means of Big Data Analytics technologies. A new path towards the world has been revealed by the big data. A an enormous ability exist in the big data with the help of several ways to save as well as to process huge substantial datasets as, at the time of processing, the large datasets are to be analysed in a required time. Regarding the Text analysis it is in its early stages also is very promising. In this thesis we will pass through censorious tools and also deal with the knowledge about the cancer study moreover we are considering some of the available visualization techniques for big data and for comparison study. We used our developed tool and FOSS. newlineIn this thesis, we also have developed Big Data Based Unsupervised Feature Learning on Deep Learning with Weighted Softmax Regression. In this sparse representation and RNN is fused in the proposed DRNNWSR (Deep Recurrent Neural Network based weighted Softmax regression) for the learning of robust feature from vast data in internet. From the feature extraction High-level structures were learned due to its rare proficiency. As a classifier the Weighted Softmax Regression (WSR) can handle large nonlinear structure of data and back propagation algorithm is utilized to provide the adjustment of softmax regression result. The entire process is made robust by the fine- tuning and the classification performance is improved and it helps to achieve fast learning rate and prevents the gradient diffusion along with the local extreme issues. When there is an increase in the dimensionality there is a decrease in the accuracy of fault ident
Pagination: A4 size paper, 142 page
Appears in Departments:Computer Sc. and Enggineering

Files in This Item:
File Description SizeFormat 
big data analytics for frequent text pattern and unsupervised feature learning.pdfAttached File4.67 MBAdobe PDFView/Open

Items in Shodhganga are protected by copyright, with all rights reserved, unless otherwise indicated.

Altmetric Badge: