Model/Technique Registration (*=required):
Information about your group/project:
Name(s) and e-mail (s) (please list primary contact first)*:
Takayuki Muranushi
Ayumi Asai
Kazunari Shibata
Shigeru Nemoto
Yuko Hada
Associated Institution/Project name/Group name*:
Kyoto University, Unit of Synergetic Studies for Space
Broadband Tower, Inc.
Website url(s):
http://www.spaceweather.kyoto/
http://54.187.234.47/prediction-result.html
Logo(s):
https://cloud.githubusercontent.com/assets/512367/7226747/632f7350-e786-11e4-85e7-9cd7c2423b3a.png
Information about your method:
Forecasting method name*:
UFCORIN (Universal Forecast Constructor by Optimized Regression of INputs)
Shorthand unique identifier for your method (methodname_version, e.g. ASSA_1, ASAP_201201):
UFCORIN_1
Short description*:
UFCORIN_1 provides a forecast for 24-hour future maximum of
solar X-ray flux, by applying neural network model using
long-short term memory (LSTM) on time-series data of 1. past
X-ray flux data and 2. wavelet features of the line-of-sight
magnetic field image data.
References:
``UFCORIN: A fully automated predictor of solar flares in GOES X-ray flux'', Space Weather (2015)
http://onlinelibrary.wiley.com/doi/10.1002/2015SW001257/full
Further Model Details:
(1)* Please specify if the forecast is human generated, human generated but model-based, model-based, or other.
It is model-based.
(2)* Does your flare prediction method forecast “M1.0-9.9” or “M and above”? Would it be difficult for you to adapt your method from one to another? Which forecast binning method do you prefer?
Our method predicts in "M and above" form. It is not difficult to switch to the other form.
(3)* How do you specify active regions in your model? Do you relate their location to other schemes such as NOAA or Catania? If so, what is your criterea to relate them?
Our scheme gives only the full disk forecast, and does not take individual active regions into account, at the moment.
(4)* Uncertainties given as an upper and lower bound is an optional field. Does your model provide uncertainties for the forecasted probability?
If yes, what percentiles do you use to determine your upper and lower bound? If you are using the XML format, multiple percentiles of the probability density distribution can be specified.
Our model does not give upper and lower bounds.
(5)* For each forecast, what prediction window(s) does your method use? E.g. Does your method predict for the next 24 hours, 48, and 72 hours or the next 24 hours, 24-48 hours, 48-72 hours? (This information is useful for displaying only comparable methods together).
Our model predicts for the next 24 hours. It is not difficult to extend the model in next 48, 72 hours, or 24-48, 48-72 hours.
(6)* For how many degrees from disk center is your full disk forecast valid for? E.g. 60 degrees; or 90 degrees (valid for entire disk).
Our model recognizes the full disk i.e. 90 degrees.
(7)* Calibration levels are optional fields. Do you have calibration for the probabilities from your model? E.g. is a 40% forecast a “high” probability for your method?
If yes, please provide the probability ranges for what is considered a “low”, “medium”, or “high” level. This can be provided in a separate file, or as part of the main template, labeled “X_Level”, “M_Level”, and “C_Level” fields which can take the values 1, 2, or 3. These levels represent probability calibration for a model for each flare class. With 1 meaning “low” probability, 2 is “medium” probability, and 3 is “high” probability.
For example in the MAG4 model:
Probabilities in range → Level
Prob < 0.02 → 1
0.02 ≤ Prob < 0.18 → 2
Prob ≥ 0.18 → 3
Our model does not produce the calibration levels.