TY - GEN
T1 - New method of remedying missing values based on support vector regression model
AU - Luo, Sen Lin
AU - Liu, Bin
AU - Pan, Li Min
AU - Ye, Ming De
AU - Ma, Zhao Yuan
AU - Zhang, Tie Mei
PY - 2010
Y1 - 2010
N2 - In the actual data there are a lot of missing values which can be properly remedied by other relevant factors in order to bring down the amount of missing information. In this paper Support Vector Regression (SVR) is applied to predict the values of abdominal circumference, body mass index and high density lipoproteins. After predicted by SVR, the average relative errors of abdominal circumference, body mass index and high density lipoproteins are respectively 4.39%, 5.73% and 11.08%, the mean absolute errors of abdominal circumference, body mass index and high density lipoproteins are respectively 3.55, 1.41 and 0.14, and the RMS errors of abdominal circumference, body mass index and high density lipoproteins are respectively 4.54, 1.8 and 0.18. Compared with other methods, the experimental results show that the mean prediction error of SVR is the smallest.
AB - In the actual data there are a lot of missing values which can be properly remedied by other relevant factors in order to bring down the amount of missing information. In this paper Support Vector Regression (SVR) is applied to predict the values of abdominal circumference, body mass index and high density lipoproteins. After predicted by SVR, the average relative errors of abdominal circumference, body mass index and high density lipoproteins are respectively 4.39%, 5.73% and 11.08%, the mean absolute errors of abdominal circumference, body mass index and high density lipoproteins are respectively 3.55, 1.41 and 0.14, and the RMS errors of abdominal circumference, body mass index and high density lipoproteins are respectively 4.54, 1.8 and 0.18. Compared with other methods, the experimental results show that the mean prediction error of SVR is the smallest.
KW - Missing values
KW - Regression model
KW - Support vector regression
UR - http://www.scopus.com/inward/record.url?scp=77953300611&partnerID=8YFLogxK
U2 - 10.1109/ICBECS.2010.5462521
DO - 10.1109/ICBECS.2010.5462521
M3 - Conference contribution
AN - SCOPUS:77953300611
SN - 9781424453153
T3 - 2010 International Conference on Biomedical Engineering and Computer Science, ICBECS 2010
BT - 2010 International Conference on Biomedical Engineering and Computer Science, ICBECS 2010
T2 - 2010 International Conference on Biomedical Engineering and Computer Science, ICBECS 2010
Y2 - 23 April 2010 through 25 April 2010
ER -