Analisis kinerja algoritma clustering dengan metode evidence accumulation

Diantari, Ratih Kirana (2014) Analisis kinerja algoritma clustering dengan metode evidence accumulation. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5110100051-Undergraduate_Thesis.pdf Download (2MB)

Abstract

Analisis cluster merupakan salah satu prosedur dalam proses penggalian data. Informasi tentang distribusi dan karakteristik data bisa diperoleh melalui proses clustering. Clustering telah digunakan dalam berbagai bidang, seperti segmentasi pasar, pengelompokan dokumen, pengenalan pola, dan analisis data spatial.Pada Tugas Akhir ini mengiimplementasikan metode clustering berbasis Consensus Clustering & Evidence Accumulation (EA). Pada metode Evidence Accumulation ada 2 parameter yg mempengaruhi kinerja algoritma clustering, yaitu k (jumlah cluster) dan t (threshold). Metode Consensus Clustering menggunakan metode K-means. Pada Tugas Akhir ini menganalisis kinerja algoritma Evidence Accumulationberdasarkan variasi parameter. Beberapa dataset yang digunakan untuk menguji kinerja metode Evidence Acumulation adalah Halfring, Spiral, dan dataset UCI. Berdasarkan hasil uji coba kedua parameter tersebut metode Evidence Accumulation mempengaruhi hasil clustering. Hasil pada metode Evidence Accumulation membentuk pola yang berbeda-beda.
==================================================================================================================================
Cluster analysis is an important data processing procedure in data mining. Through clustering, valuable information such as data distribution and characteristics can be acquired. Clustering has been widely applied in various fields, such as market segmentation, document clustering, pattern recognition, and spatial data analysis. In this final project, clustering method which is used are Consensus Clustering and Evidence Accumulation (EA). There are two parameters of EA that affect the performance of clustering algorithm, namely k (number of clusters) and t (threshold). Consensus clustering method used a K-means clustering method. The purpose of this final project to analyze performance of EA algorithms based on parameter variations.Some datasets used to test the performance of the EA method are Halfring, Spiral, and the UCI datasets. Based on the results of the both test parameters, EA methods affect the results of clustering. The results of the EA methods form different patterns.

Item Type:	Thesis (Other)
Additional Information:	RSIf 005.74 Dia a-2014
Uncontrolled Keywords:	analisis cluster, consensus clustering, k-means clustering, evidence accumulation, cluster analysis, consensus clustering, k-means clustering, evidence accumulation.
Subjects:	Q Science > QA Mathematics > QA76.758 Software engineering
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Informatics Engineering > 55201-(S1) Undergraduate Thesis
Depositing User:	Mr. Marsudiyana -
Date Deposited:	23 Oct 2023 08:08
Last Modified:	23 Oct 2023 08:08
URI:	http://repository.its.ac.id/id/eprint/104873

Actions (login required)

View Item