Authors: Rahmad Kurniawan, Okfalisa, MZA Nazri
Publication date: 2014/10/21
Conference: International Conference on Science and Technology for Sustainability 2014
Issue: October 21-22, 2014
Publisher: Institute of Electrical and Electronics Engineers, UIN Sultan Syarif Kasim Riau , Proceeding, Volume 1, October 2014
Abstract— Decision Tree (DT), Artificial Neural Network (ANN), and Rough Set Theory (RST) are commonly techniques of classification in data mining. This study proposed a comparative analysis of the above techniques by using open source's software ROSETTA and WEKA on the case study of housing dataset. The analysis examined the effectiveness of the techniques according to ‘no free lunch’ theorem. Herein, WEKA as open source software which consists of a collection of machine learning algorithms is applied for data mining tasks. Meanwhile, ROSETTA is as a toolkit for pattern recognition and data mining within the framework of rough set theory. The parameters to evaluate the performance of each classification modeling are including accuracy; amount and lenght of the rules for each model; standard deviation and error rate. Based on 9 experimental models, ANN provided the highest accuracy amongst the techniques 82.24% and the lower error rate. Thus, it can be concluded that the lack of a technique can be accommodated by the advantages of another techniques as can be learned from the ‘no free lunch’ theorem.
Keywords—artificial neural network; decision tree; housing dataset; rough set theory;
Artikel keren lainnya:
1 Tanggapan untuk "A Comparative Study Of Decision Tree, Artificial Neural Network And Rough Set Theory: Case Study On Housing Dataset"
pak, ingin tanya tentang teknik data mining untuk memprediksi penjualan suatu produk, model yang umum untuk memprediksi data historis time series penjualan yang berbentuk seasonal trend selain holt-winter apa ya pak? saya pernah menulis artikel tentang holt winter berikut: http://datacomlink.blogspot.co.id/2015/12/serumit-apa-forecast-metode-holt.html
Post a Comment