Score sheet details

In this sheet, you can examine various accuracy measures for each category that is specified in the training data file. The system generates this sheet for each algorithm in the analysis.

The score sheet contains the following types of data for each record:

• Category - It indicates categories or sentiments that you specified in the training data file in the Results column.
• Manual occurrence count - It indicates the total number of times when you identified a given category or sentiment as the predicted outcome of the analysis (that is, the manual outcome).
• Machine occurrence count - It is the sum of true positives and false positives returned by the algorithm. The value indicates the total number of times when a given category or sentiment was indicated as the outcome with the highest accuracy score (that is, the machine outcome).
• True positive - It indicates the number of outcomes in a given category that the algorithm classified correctly. An outcome is considered a true positive when the manual outcome matches the machine outcome.
• Precision - It is the number of true positives divided by the number of machine outcomes. This score indicates how precise the algorithm is at retuning the true positive results with respect to the outcomes that were returned by the algorithm (that is, the outcomes with the highest accuracy score).
• Recall - It is the number of true positives divided by the number of manual outcomes. The score indicates how precise the algorithm is at returning the true positive results with respect to the outcomes that should have been returned by the algorithm (that is, the outcomes you predicted).
• Fscore - It is the measure of the accuracy of the text analysis model. The score is the weighted average of the precision and recall. The Fscore values range from 0 to 1.

Open topic with navigation