I am currently solving one classification problem in which there are 235 rows and 51 columns.I want to know what percentage of rows I should use for training of my model and what percentage I should use for testing of my model for better accuracy
@hinduja1234 - The percentages of observation kept for testing depends on a total number of observations available. For a large dataset, you can keep 40% for testing and for small dataset even 10% can do the job. Idea is to keep enough number of observation to train your model and perform testing simultaneously.
Here you can keep out 10% of data for testing your model because the number of rows is not so large.