TY - GEN
T1 - Point and interval estimation method for auto-regressive model with nonnormal error
AU - Lim, Bo Mi
AU - Kim, Jongwoo
AU - Kim, Sung Shick
AU - Baek, Jun Geol
PY - 2013
Y1 - 2013
N2 - Estimation in time series analysis aids in making a reasonable decision by providing a value for point estimation and a range of interval estimation. An auto-regressive model is designed for the time series analysis. However, the auto-regressive model may cause decreasing accuracy and prediction in estimating parameters because it uses the assumption that the distribution of error term follows a normal distribution. In reality, there are plenty of data indicating that the distribution of error term does not follow the normal distribution. Thus, we propose a method for solving this problem by using a Pearson distribution system and maximum likelihood estimation. Compared with existing methods, the proposed method can be applied to various time series data requiring high accuracy and prediction.
AB - Estimation in time series analysis aids in making a reasonable decision by providing a value for point estimation and a range of interval estimation. An auto-regressive model is designed for the time series analysis. However, the auto-regressive model may cause decreasing accuracy and prediction in estimating parameters because it uses the assumption that the distribution of error term follows a normal distribution. In reality, there are plenty of data indicating that the distribution of error term does not follow the normal distribution. Thus, we propose a method for solving this problem by using a Pearson distribution system and maximum likelihood estimation. Compared with existing methods, the proposed method can be applied to various time series data requiring high accuracy and prediction.
KW - Auto-Regressive Model
KW - Interval Estimation
KW - Maximum Likelihood Estimation
KW - Nonnormal data
KW - Pearson Distribution System
KW - Point Estimation
UR - http://www.scopus.com/inward/record.url?scp=84885996050&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84885996050&partnerID=8YFLogxK
U2 - 10.1109/BigData.Congress.2013.57
DO - 10.1109/BigData.Congress.2013.57
M3 - Conference contribution
AN - SCOPUS:84885996050
SN - 9780768550060
T3 - Proceedings - 2013 IEEE International Congress on Big Data, BigData 2013
SP - 379
EP - 386
BT - Proceedings - 2013 IEEE International Congress on Big Data, BigData 2013
T2 - 2013 IEEE International Congress on Big Data, BigData 2013
Y2 - 27 June 2013 through 2 July 2013
ER -
