Please use this identifier to cite or link to this item:
http://hdl.handle.net/10603/267611
Title: | Design and development of clustering algorithm to improvise computation of data stream based on outlier attributes |
Researcher: | Shukla, M.S. |
Guide(s): | Kosta, Y.P. |
Keywords: | Clustering Engineering and Technology,Computer Science,Computer Science Information Systems Non Stationary Systems Outlier Attributes Stream Data Mining |
University: | RK University |
Completed Date: | 23/09/2019 |
Abstract: | quotIn recent years, advancement in hardware technology has allowed us to automatically record transaction of everyday life at rapid rate. Such process/systems lead to large amounts of data, which grow at an unlimited rate. These data instances are referred as streams of data. Systems that generate this type of data are Sensor networks, Electricity generation, Telecommunication systems, Real time surveillance system etc. Generation and collection of such data gives rise to the need of intelligent processing of data which in turn can help in analyzing the data and thus aid in decision making process too. newlineCrucial challenges faced while dealing with stream data is its sheer volume and speed at which it is generated. Other characteristics like Concept Evolution, Feature Evolution of incoming data and its continuous arrival add more complication in correct analysis, henceforth mining stream data is extremely tedious task. Moreover, scenario for unsupervised learning is even complicated, as without class label of stream data, classifying it into a group (clustering) seems impractical. Clustering is a unique example of one such process. Clustering of Stream data means analyzing streams and extracting valuable patterns from them and then making cluster based on the patterns. Clustering finds its importance in today s trend because always it is not possible have data which is labeled and making cluster from the gathered data can help a lot in quick decision making about category of data and its resultant consequences . newlineDespite of being an interesting area of research, due to challenges in mining knowledge from stream data less work is observed in this field. Traditional algorithms exist in data mining but as these algorithms are multi-pass algorithm so, they are not suited for stream data mining also these algorithms are not able to detect and deal with concept shift in stream data, which is its one of the special features. Keeping this in mind, we have performed analysis and evolution of existing stream data mining algorith |
Pagination: | - |
URI: | http://hdl.handle.net/10603/267611 |
Appears in Departments: | Faculty of Technology |
Files in This Item:
File | Description | Size | Format | |
---|---|---|---|---|
01_cover page.pdf | Attached File | 49.76 kB | Adobe PDF | View/Open |
02_certificate.pdf | 81.19 kB | Adobe PDF | View/Open | |
03_declaration.pdf | 95.86 kB | Adobe PDF | View/Open | |
04_acknowledgments.pdf | 39.63 kB | Adobe PDF | View/Open | |
05_table of contents.pdf | 79.03 kB | Adobe PDF | View/Open | |
06_list of tables.pdf | 99.2 kB | Adobe PDF | View/Open | |
07_list of figures.pdf | 143.58 kB | Adobe PDF | View/Open | |
08_list of abbreviations.pdf | 67.75 kB | Adobe PDF | View/Open | |
09_abstract.pdf | 114.4 kB | Adobe PDF | View/Open | |
10_graphical abstract.pdf | 155.23 kB | Adobe PDF | View/Open | |
11_chapter 1.pdf | 257.71 kB | Adobe PDF | View/Open | |
12_chapter 2.pdf | 520.21 kB | Adobe PDF | View/Open | |
13_chapter 3.pdf | 885.85 kB | Adobe PDF | View/Open | |
14_chapter 4.pdf | 257.86 kB | Adobe PDF | View/Open | |
15_chapter 5.pdf | 746.78 kB | Adobe PDF | View/Open | |
16_references.pdf | 127.43 kB | Adobe PDF | View/Open | |
17_list of publication.pdf | 187.85 kB | Adobe PDF | View/Open |
Items in Shodhganga are licensed under Creative Commons Licence Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA 4.0).
Altmetric Badge: