Expected Behaviour
After splitting data with Percent feature, following thing will happen:
1).User enters the split percentage.
2). Candis firstly randomize the provided data.
3). Then split the data into training and testing dataset with given percentage/ratio.
3). Both datasets gets saved in the user data directory.
4). Training dataset will then be trained similarly how it is currently done on Candis.
5). Then user need to test against the testing dataset using yet to be implemented Predict option on candis.
Actual Behaviour
Currently what is happening, user provides data, candis split the data itself using K-Folds technique, runs the pipeline, and give the model accuracy against the data/CEL provided by the user.
Expected Behaviour
After splitting data with Percent feature, following thing will happen:
1).User enters the split percentage.
2). Candis firstly randomize the provided data.
3). Then split the data into training and testing dataset with given percentage/ratio.
3). Both datasets gets saved in the user data directory.
4). Training dataset will then be trained similarly how it is currently done on Candis.
5). Then user need to test against the testing dataset using yet to be implemented Predict option on candis.
Actual Behaviour
Currently what is happening, user provides data, candis split the data itself using K-Folds technique, runs the pipeline, and give the model accuracy against the data/CEL provided by the user.