- Most learning methods use a tolerance threshold EPSILON=
. Before using these methods, it is better to normalize the data between 0 and 1, if their range is very high or very small.
- The HFP induction method is based on a data file only. It considers that there is a single output which corresponds to the data file last column.
- Once memory is allocated, some Java virtual machines (JVM) do not free it while the Java process is active, even if the memory is no longer used. Large arrays can be allocated in FisPro when reading big data files. The only solution to free memory is then to save current work, exit FisPro and relaunch it.
