Embed this Script in your web site
|Created||Wed, 17 May 2017 22:51:17 +0000|
|Published||Fri, 19 May 2017 12:58:22 +0000|
This script takes two datasets, a training and a holdout, a list of predictors to be built, and a list of parameters to pass to them. (By default they are a model, a bagging ensemble, a random decision forest, a boosted ensemble, and a logistic regression.) The script then runs a batch prediction on the holdout dataset for each predictor, and finally appends the holdout dataset with the prediction of each predictor and the most popular prediction over all.
A training dataset to build the predictors
A holdout dataset to predict
A list of predictors to build. Possible values include "model", "ensemble", "logisticregression", and "lr". The type of ensemble created is dependent on the parameters passed by the parameters list. The default is a model, a bagging ensemble, a random decision forest, a boosted ensemble, and a logistic regression.
A list of parameters to pass to your predictors. The default is the default BigML settings for the default predictors above
- A copy of the holdout dataset appended with all the predictions and the most popular prediction last