a classification template in R to accelerate data preparation and modeling
This classification template in R provides a fast and easy way to prepare your data and build and compare different classification models. The sole purpose of this template is to help you to build a model fast and easy.
The template contains three steps:
- Data Preparation: all steps needed to prepare your data for using it in analyses or models and creating a training and test set. For instance, the processing of outliers, missings, non-zero variance, correlation, optimal binning and even feature selection (optional);
- Modeling: the template automatically tries to create different types of models. By default, the template uses Extreme Gradient Boosting, Neural Networks, Logistic Regression and Ensemble Methods, but this is easy to adapt to your preferences;
- Evaluation: visual insights into the model(s), its predictors and its performance compared to the other models.
Together with this template, we deliver our own R package. The sole purpose of this package is to make your life easier when it comes to data preparation and modeling. We have incorporated many difficult and time-intensive matters into this package, such as loading data from different sources with one function, or resolving all missings or outliers at one fell swoop.
gain insights from predictive models faster and easier
- Save time and effort: let your analysts be more effective and efficient, building a new model doesn’t take days anymore, but just (a few) hours;
- Immediately get good and trustworthy models;
- The template is easy to adapt to your own needs;
- The template a good starting point for the development of models in R;
- The template is most suitable to teach your people to work with R.
every analyst who wants to use or uses R for predictive modeling
Our classification template is a must-have for all analysts, whether they are already working in R or want to start doing that. It provides a fast, easy and stable way to prepare your data and to build and compare multiple models.