Electronic message text classification framework selection

A system comprising:

• a database containing a number of text classification frameworks which classify incoming electronic messages;
• a classification scorer to score classifications of multiple text classification frameworks;
• a simulator to compare a probability that re-training a current text classification framework improves performance with a probability that switching to a different text classification framework improves performance;
• and
• a framework controller to select one of the multiple text classification frameworks based on an output of the classification scorer and an output of the simulator;
• wherein the multiple text classification frameworks comprise the current text classification framework and other text classification frameworks.

A computer program product for selecting a text classification framework, the computer program product comprising a computer readable storage medium having program instructions embodied therewith, the program instructions executable by a processor to cause the processor to:

• classify an incoming electronic message using a current text classification framework;
• score the classification of the current text classification framework;
• perform a simulation to compare a cost of re-training the current text classification framework to improve performance with a cost of switching to a different text classification framework to improve performance;
• select one of the multiple text classification frameworks, which comprises the current text classification framework and other text classification frameworks, for text classification based on:
• the score of the classification;
• and
• an output of the simulation;
• when the different classification framework is selected:
• prevent execution of the current text classification framework;
• and
• call the different text classification framework to classify the electronic message;
• and
• when the current classification framework is selected:
• prevent execution of the different text classification framework;
• and
• determine whether to re-train the current text classification framework to improve performance.