Please use this identifier to cite or link to this item: http://archives.univ-biskra.dz/handle/123456789/13213
Title: Selecting SOA design patterns using machine learning techniques
Other Titles: informatique
Authors: thamer, halima
Issue Date: 20-Jun-2019
Abstract: with the continuously increasing amount of textual informations, there is a pressing need to structure them. Text classification (TC) is a technique which classifies textual information into a predefined set of categories where documents can be automatically classified based on their contents. Automatic classification of data is one of the main applications of machine learning algorithms. This master’s thesis explores a way in which text classification is used to categorize user (developer) problems using machine learning algorithms so we can help him to find the appropriat design pattern knowen,that finding and selecting a suitable soa design pattern has always been a challenging task especially for young developer during the designing phase because usually, a designer has to consult and search extensively to find a suitable pattern. Among many patterns, few are considered to be relevant to solve a problem We can define a 'suitable Soa pattern' as a solution, which is aligned with the developer's problem and has a good affect. With the help of already manually classified data that’s the set of SOA Patterns a model can be learned using learning teckniques like (naïve-Bayesian, k-nearest-neighbor, svm, decision-Tree). So that we can find the class of design patterns that our problem belongs to. then we use the patterns of this class as dataset to find the most appropriate one or in other words the closest design pattern to the user problem using cosine similarity. . Keywords: classification; cosine similarity; SOA Patterns.
URI: http://archives.univ-biskra.dz/handle/123456789/13213
Appears in Collections:Faculté des Sciences Exactes et des Science de la Nature et de la vie (FSESNV)

Files in This Item:
File Description SizeFormat 
thamer_halima.pdf1,93 MBAdobe PDFView/Open


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