Hi.

the following paper provides a very comprehensive survey of typical performance/evaluation measures typical in classification:

http://rali.iro.umontreal.ca/rali/sites/default/files/publis/SokolovaLapalme-JIPM09.pdf

It comprises for instance, Average Accuracy, Error Rate, Precision, Recall, F-score,

Precision, Recall. And many more

Hi.

the following paper provides a very comprehensive survey of typical performance/evaluation measures typical in classification:

http://rali.iro.umontreal.ca/rali/sites/default/files/publis/SokolovaLapalme-JIPM09.pdf

It comprises for instance, Average Accuracy, Error Rate, Precision, Recall, F-score,

Precision, Recall. And many more

You can use Support Vector Machines and diagnostic tools as ROC, area under curve to see the discrimination ability of your model.

true positives, false positive, true negatives, false negatives are converted to precision, recall, accuracy. Then F-Measure is used. You can use cross validation technique as well.

Maybe you can use the Rnx function which represent the improvment of a technique compared to a random distribution. As Arnau S-C said, it depends of your data structure, and also what you want to do with it.

TP rate, FP rate, recall precision, accuracy, ROC curve, error rates are some of the measures to judge the performance of classification...

paper of Marco F. Huber is very useful...thanks for sharing...

For complex structure or phenomena such as complex networks, the effective way of verifying is to examine whether there are far more small classes than large ones within the complex structure:

Jiang B. and Ma D. (2015), Defining least community as a homogeneous group in complex networks, Physica A: Statistical Mechanics and its Applications, 428, 154-160.

https://www.researchgate.net/publication/271771290_Defining_Least_Community_as_a_Homogeneous_Group_in_Complex_Networks?ev=prf_pub

Article Defining Least Community as a Homogeneous Group in Complex Networks