An Analysis on Two Different Data Sets by using Ensemble of k-Nearest Neighbor Classifiers

Authors: , ,

Abstract: An ensemble of classifiers is proposed in this study where we combined the k-nearest neighbor classifier with LAD Tree through stacking. Two different data sets which are macroeconomic and the International Risk Country Risk are used for application. Both of the data are taken for 27 countries from the first quarter of 1984 until the fourth quarter of 2011. Before we applied those two data sets on the ensemble of classifiers, we have done the multicollinearity test to see if the predictor variables are highly correlated. By using the variance inflation factor, the results showed that some predictor variables are highly correlated in the macroeconomic data set. To solve the multicollinearity problem, we used principal component analysis that is available in SPSS as it can reduce a set of factors. The remaining variables are used in the analysis of the proposed method to observe the prediction accuracy. It is proven that different type of data does affect the prediction accuracy, but then it is different for every country.

WSEAS Transactions on Mathematics, ISSN / E-ISSN: 1109-2769 / 2224-2880, Volume 13, 2014, Art. #76

PDF
XML
Cite
×
Citation Tools
WSEAS AMA
Nor Azuana Ramli, Mohd Tahir Ismail, Hooy Chee Wooi, "An Analysis on Two Different Data Sets by using Ensemble of k-Nearest Neighbor Classifiers," WSEAS Transactions on Mathematics, vol. 13, pp. 780-789, 2014, DOI:
Nor Azuana Ramli, Mohd Tahir Ismail, Hooy Chee Wooi. An Analysis on Two Different Data Sets by using Ensemble of k-Nearest Neighbor Classifiers. WSEAS Transactions on Mathematics. 2014;13:780-789.
Copy to Clipboard
Citation copied to Clipboard

Certification

---