PENGARUH PERHITUNGAN RENTANG DATA TERHADAP PERFORMANSI ALGORITMA CLUSTERING SELF ORGANIZING MAP
Total View This Week0

Institusion

Universitas Sriwijaya

Author

TRI KURNIA SARI (STUDENT ID : 09021281520105)
Dian Palupi Rini (LECTURER ID : 0023027804)
Novi Yusliani (LECTURER ID : 0008118205)

Subject

R858-859.7 Computer applications to medicine. Medical informatics 

Datestamp

2019-10-02 02:05:18 

Abstract :
Clustering is a valuable research field in data mining. Clustering groups data based on its proximity or similarity. An important component in clustering is the distance measurement between data points. The difference in distance measurement method of data will affect how close the distance between the data. Therefore, this research will examine the effect of distance measurement of data on the performance of clustering algorithm. This study uses Self Organizing Map algorithm. There are three distance measurement method are Euclidean distance, Manhattan distance, and Chebyshev distance. This research uses 4 data set are Iris, Wine, Car Evaluation, and Abalone. The cluster will be evaluated using Davies Bouldin Index. Self Organizing Map and Chebyshev distance produce better cluster than Euclidean distance and Manhattan distance. However, the computation time is longer than Euclidean distance and Manhattan distance. 

File :
RAMA_55201_09021281520105_0023027804_0008118205_01_front_ref.pdf
RAMA_55201_09021281520105_0023027804_0008118205_02.pdf
RAMA_55201_09021281520105_0023027804_0008118205_03.pdf
RAMA_55201_09021281520105_0023027804_0008118205_04.pdf
RAMA_55201_09021281520105_0023027804_0008118205_05.pdf
RAMA_55201_09021281520105_0023027804_0008118205_06.pdf
RAMA_55201_09021281520105_0023027804_0008118205_07_ref.pdf
RAMA_55201_09021281520105_0023027804_0008118205_08_lamp.pdf