The Random Forest Classification (RFC) and Random Forest Regression (RFR) in Jasp have not K-fold option in training data menu and resuts added in data (column) is only the resuts of classiication or regression, without any information of probabiity (classification) or statistic metrcs as mean, standard deviaiton, median and 5th and 95th centile for regression. For example, I run a RFC or RFR with 1000 trees. The expected result column added in dataset is the classification (A, B, C, so on) or a continuous value for RFR. But, by producing 1000 trees, we have in fact 1000 results by point. So, we could extract the frequency (or probability) of a classification, or the devation statistics for a regression, added in the dataset. In an ideal world, when we use the produce model into the "prediction menu", it will be great to have a column with the result (classe or continuous value) and the uncertainty measurement (probability or classe and std deviation, median, centiles fot regression result)
Purpose
Improve the results by adding incertainty metrics and add a K-fold option
Use-case
machine learning process
Is your feature request related to a problem?
lack of k-fold option and uncertainty added in results
Is your feature request related to a JASP module?
Machine Learning
Describe the solution you would like
K-fold option in RFR and RFC in validation and training data // added columns in dataset with result (existing now) an uncertainty resuts (probability for classes, std deviation, median and centiles for regression)
Description
The Random Forest Classification (RFC) and Random Forest Regression (RFR) in Jasp have not K-fold option in training data menu and resuts added in data (column) is only the resuts of classiication or regression, without any information of probabiity (classification) or statistic metrcs as mean, standard deviaiton, median and 5th and 95th centile for regression. For example, I run a RFC or RFR with 1000 trees. The expected result column added in dataset is the classification (A, B, C, so on) or a continuous value for RFR. But, by producing 1000 trees, we have in fact 1000 results by point. So, we could extract the frequency (or probability) of a classification, or the devation statistics for a regression, added in the dataset. In an ideal world, when we use the produce model into the "prediction menu", it will be great to have a column with the result (classe or continuous value) and the uncertainty measurement (probability or classe and std deviation, median, centiles fot regression result)
Purpose
Improve the results by adding incertainty metrics and add a K-fold option
Use-case
machine learning process
Is your feature request related to a problem?
lack of k-fold option and uncertainty added in results
Is your feature request related to a JASP module?
Machine Learning
Describe the solution you would like
K-fold option in RFR and RFC in validation and training data // added columns in dataset with result (existing now) an uncertainty resuts (probability for classes, std deviation, median and centiles for regression)
Describe alternatives that you have considered
No response
Additional context
No response