Data

The time series for EFchallenge 2018 will be uploaded on March 5, 2018
For an impression you can find the time series sets of EFchallenge 2017 below. Note that this years edition will only have one round with 5 simulated and 5 real time series.

Your forecasts will be evaluated with respect to the conditional expectation forecast in RMSE sense per time series. Points are awarded based on the ranking for each of the five time series. The group with the lowest RMSE recieves 10 points, second lowest 8 etc. The fifth group and lower recieve 0 points. The winners for the entire competition are determined by summing all the obtained points for all time series.

Submission form
Form
Deadline EFchallenge 2018
Final: May 20, 2018 at 23.59

Data EFchallenge 2017 example
For the third round, you are asked to make a ten step ahead forecast for the following five time series (please note that the first 500 observations are equal to the time series from round 2):

TS1
TS2
TS3
TS4
TS5

For the second round, you are asked to make a ten step ahead forecast for the following five time series (please note that the first 250 observations are equal to the time series from round 1):

TS1
TS2
TS3
TS4
TS5

For the first round, you are asked to make a ten step ahead forecast for the following five time series:

TS1
TS2
TS3
TS4
TS5

Please note that the files above are *.csv files and that the EFchallenge uses a point as decimal mark!