On mathematical analysis of clustering algorithms in text mining
Is this your thesis?
This record came from a bulk archive import. If it’s yours, link it to your profile.
Abstract (EN)
Text mining is an interdisciplinary research field enabling multi-dimensional data analysis including methods such as classification, clustering and feature extraction. The purpose of clustering problem, included in this field, is to sub-cluster data with similar features with each other. In cases where handled data are composed of texts, in order to apply clustering methods, documents should firstly be converted to digital data. Creating feature matrixes of each document, they are made applicable to clustering algorithms. When data size is too large, run-time of algorithms can be too long. Therefore, it is expected from algorithms to provide optimum solutions in a shorter time by making rights choices. In this thesis, the k-means based clustering approach has been used for solving the text clustering problem. The mathematical analysis of clustering problem was done by selecting starting central points of the handled k-means clustering algorithm, with various methods. Distance function of each algorithm was calculated and compared to each other. Calculation tests on real data sets appear that various starting central points lead to how different consequences.
Author
Oğuzcan Uludağ
How to Cite
Oğuzcan Uludağ (Master Thesis). On mathematical analysis of clustering algorithms in text mining, 2017, Ege University.
Keywords
License
Tüm Hakları Saklıdır
This work is shared under the specified license terms.
More theses from Ege University
- Increasing the fertilization yield of trout eggs(2013)
- Effect of microfibers on mechanical and durability properties of cement mortar mixtures(2015)
- Comparative lexicon of Qipchaq Turkish on the base of codex cumanicus(2015)
- Investigation of phyto-microbial remediation of arsenate resistance in Gönen rice by removal of arsenite(2019)
- The use of social media in city branding and its contribution to brand image: Case of İzmir(2019)
- Effect of using stevia extract as natural sweetener during production of ice cream with low calorie on quality criteria of product(2015)
