Performance Evaluation Metrics Clausole campione

Performance Evaluation Metrics. To evaluate the performance of the algorithms the following quan- tities were defined (▇▇▇▇▇▇▇▇▇▇ et al., 2010): - tp (True Positive), the number of images correctly classified in the class of belonging; - tn (True Negative), the number of images correctly classified in another class; - fn (False Negative), the number of images incorrectly clas- sified in other classes; - fp (False Positive), the number of images incorrectly classi- fied in the class; We have employed the following metrics to compare the perfor- ▇▇▇▇▇ of the algorithms that used for the evaluation of the image sentiment: