| |
| | [No title] |
 | | While WekaClassify was developed by Satanjeev Banerjee as a part of that package, we have found it generally useful and are distributing it separately as well, in the hopes that it will be useful for users of the machine learning suite Weka, regardless of their application area. |
 | | WekaClassify carries out classification based on a previously learned model by Weka, and produces output such that each possible answer is "scored" based on whatever criteria the learned model might use (confidence scores, probabilities, etc.). |
 | | This model has to be created by training some classifier (such as a decision tree, a neural network, a naive Bayesian classifier, etc) on training data consisting of instances of the same target word as that in the test arff file. |
| www.d.umn.edu /~tpederse/Code/Readme.WekaClassify-0.4.txt (777 words) |
|