machine learning - What size should my data set be used for classification experiment? -
Working for some, let me compare some classification techniques (vector vector machine, neural network, decision tree, etc.) Will happen. My contact person at the university asked me to use the Kagal Data Set.
The Data Set includes a set of training sets of 150,000 borrowers and 100,000 borrows for me, only training sets are useful, because the test set is not the result of borrowers.
My question is, how many examples should I use, keeping in mind the computational endeavor of a large data set
- How many examples would you use?
Data split, training at 90% and testing at the remaining 10%:
size = int (lane (brown_tagged_seant) * 0.9) size 4160
Comments
Post a Comment