I'm running liblinear on text classification using -s 0 and -s 6 in order to get probability estimates of a multiclass classification task.
I read through the guides, sites, documents, feature scaling etc but the accuracy of each classifier is always low ? e.g. 14%
How do I further improve the accuracy of the classifier?
Any help most useful
I'm running liblinear on text classification using -s 0 and -s 6 in order to get probability estimates of a multiclass classification task.
I read through the guides, sites, documents, feature scaling etc but the accuracy of each classifier is always low ? e.g. 14%
How do I further improve the accuracy of the classifier?
Any help most useful