accuracy, precision, recall, and the f1-score stands as a good measure in the classification algorithms, but I would to compare two neural code represented with an unsupervised method like K-Means etc, how I can do it?
Here is the link on metric for clustering performance evaluation on sklearn:
https://scikit-learn.org/stable/modules/clustering.html#clustering-performance-evaluation
between_SS/total_SS a measure of the goodness of the classification k-means. Ideally you want a clustering that has the properties of internal cohesion and external separation, i.e. the BSS/TSS ratio should approach 1.
you can use miimim intra clustering as the mine internal measure also the result can be evaluated using CH, Duun etc, while f1-score and entropy a good measure for external measure.
You can use minimum mean square error, F-measure, entropy for unsupervised learning task
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Graphs