Does it have some work of general amount of faulty data in WSN (Wireless Sensor Network) or any kind of sensors?
I would like to generate some sensor's faulty data for machine learning training, but I don't know at which amount I should stop (25%,50%,...) ?
Mohamed Seifeddine Azzabi
unfortunately there's no clear response to this question, specially for unsupervised algorithms.
Its depends on situation but when it comes to practice, 1% seems to be the adequate rate.
0 votes
0 thanks
- Badges
- Science topic
- Similar topics
- Correction
More Mohamed Seifeddine Azzabi's questions See All
Similar questions and discussions