Current likelihood calculation multiplies probabilities (whose values are between 0 and 1) thus having longish docs with unfrequent words for some class/category may lead to multiple double multiplications to return 0 even if that's not the correct value (thus assigning such a class 0 probability too).
Probably using loglikelihood and/or BigDecimals may help.
|Transition||Time In Source Status||Execution Times||Last Executer||Last Execution Date|
|227d 18h 12m||1||Tommaso Teofili||25/Nov/13 08:10|
|111d 4h 51m||1||David Smiley||16/Mar/14 13:02|
|Status||Resolved [ 5 ]||Closed [ 6 ]|
|Status||Open [ 1 ]||Resolved [ 5 ]|
|Resolution||Fixed [ 1 ]|
|Field||Original Value||New Value|
|Fix Version/s||4.7 [ 12325572 ]|