About: This article introduces the Model node, a Special node within Construct.
Location: Node panel
Table of Contents
Feature Overview
The Model node is used to perform the mathematical computations necessary to build a statistical model predicting the behavior of a chosen Y Variable. The output model is a mathematical formula that is comprised of one or more of the other variables found in the dataset entering the Model node.
Note: For more advanced statistical modeling, users can bring their blended and cleaned datasets into Predict - Rapid Insight's automated predictive modeling and advanced analytics product. To learn more about Predict, please reach out to RI-Support@eab.com.
Model Types
The Model node will determine which type of model to build based off the designated Y Variable.
Logistic Regression
Logistic regression is used when the Y Variable is a binary. The resulting model looks to predict the likelihood of an event to take place.
OLS Regression
Ordinary Least Squares (OLS) Regression is used when the Y Variable is continuous. Users may select the Logistic regression method to model continuous variables, though this is not recommended. The resulting model looks to predict the rate or quantity of an event.
Build Model Now
The easiest and fastest way to build a model is by clicking the Build Model Now button. This option allows Construct to build the best model that can build based on the current dataset. It takes all non-linearities in the data into account and makes all of the appropriate transformations on the candidate variables automatically.
Model Information
The Model Information window contains all of the variables available when building a model. It also displays the model’s variable coefficients, standard errors, t-values, and p-values.
Comments
0 comments
Article is closed for comments.