TY  - JOUR
T1  - Performance Evaluation of Statistical and Artificial Neural Network based Short Term Load Forecasting Techniques
AU - , Ajay Shekhar Pandey AU - , D. Singh AU - , S.K.Sinha AU - , S.P. Singh 
JO  - International Journal of Electrical and Power Engineering
VL  - 2
IS  - 6
SP  - 393
EP  - 397
PY  - 2008
DA  - 2001/08/19
SN  - 1990-7958
DO  - ijepe.2008.393.397
UR  - https://makhillpublications.co/view-article.php?doi=ijepe.2008.393.397
KW  - Short term load forecasting
KW  -radial basis function
KW  -multiple linear regression
KW  -self-organizing
KW  -numerical taxonomy
KW  -back propagation
AB  - The performance evaluation of various short term load forecasting approaches has been made in this paper. The methods are selected to reflect the different categories of load forecasting approaches. These methods are Multiple Linear Regression and Time Series in Statistical/ Conventional approach, Feed Forward Neural Network in supervised Artificial Neural Network approach, Radial Basis Function Neural Network in Unsupervised/Supervised category and Numerical Taxonomy method in Self Organizing category. The performance is studied on the historical data of a Canadian utility. It is observed that the neural network based methods are quite accurate as compared to conventional statistical methods. Statistical methods are accurate only when the load behaviour is less erratic. The self-organizing Numerical Taxonomy method shows the best results.
ER  -