2014
---
2
Student name:
Ilana Gitlins
Department: Industrial engineering
andmanagement
Advisor name:
Dr. Yevgenia Apartsin
Dataminingmethods for prediction
of electricity loads
Prototype flowdiagram:
Theprototype andmodels are implemented using
Matlab
Project results:
The final error in forecasting for all the zone is calculate by the average MAPE of
the forecasted values from the best fitted model for each zone. The average
accuracy level
for the entire dataset is
91.9%
Accuracy level
91.9%
Data loading
Data
preprocessing
Input andmodel
preparation
Models testing
Conclusions
•
Load loads
•
Load
temperature
•
Load
benchmark
•
Choose Sample
•
Divide training
and test sets
•
Select
parameters
•
Testmodels
on training
set
•
Testmodels
on test set
•
Compare
errors
results
•
Choose the
modelwith
lowesterror
•
Remove
nulls
•
Interpolate
outliers
•
Add
parameters
