MBF Summary

1
MBF Summary

• The algorithm
• Feature extraction
• Retrieval
• Forecast

2
MBF Experiment results

• A) Query sequence
• B) First candidate
• retrieved
• C) Actual weather
• D) Forecast

3
MBF - Discussion

• Retrieval depends heavily on the available
  database.
• Amount of stored data and forecast accuracy
  relation must be investigated.
• Relation between forecast lead time ? and
  matching length L must be explored.
• Can we use DB from other places ?

4
MBF - Enhancements

• Combine several retrieved sequences instead of
  selecting just one.
• More accurate forecast can be achieved by
  combining the system with physical atmospheric
  information, e.g. temperature, pressure and
  humidity.
• A man-machine interface is needed.

5
MBF The Dream Team..
S. Suzuki
K. Otsuka
T. Horikoshi
H. Kojima