TomasB
Reputation: 744
Different attack types (labels) in training set than in test set
I am currently working on KDD CUP 99 and NSL-KDD datasets (Intrusion detection datasets). They are divided into training and test sets, however when I go through the sets they have different attack types (data labels).
Is it normal?
The accuracy of predictions on the test set will be decreased because of the classes which my classifier was not be able to classify correctly since it was never trained on them.
Thank you!
Upvotes: 0
Views: 48
Answers (0)
Related Questions
- Weka: Incompatible Training and Test Sets
- Anomaly detection with machine learning without labels
- High precision recall for train data but very poor for test data in classification problem
- Deep Learning: Multiclass Classification with same amount of labels between the training dataset and test dataset
- Training Data Vs. Test Data
- ANN different results for same train-test sets
- WEKA - Classification - Training and Test Set
- KDD1999 dataset Features exolaination
- Neural network and IDS
- Feasibility of Machine Learning techniques for Network Intrusion Detection