Building A New Model

Learn how to build a new predictive model.

Forwrd Support avatar
Written by Forwrd Support
Updated over a week ago

Prerequisites

In order to be able to build a new model, you should have completed the following steps:

  1. Integrate Forwrd to the relevant data sources

  2. Define the model's metrics

Introduction

Now that you've integrated Forwrd with your data sources, created a base for the model, and defined the metrics, you are ready to build your new predictive model. Forwrd scores your data using four buckets: Excellent, Good, Fair, and Poor. Excellent means that this record is very likely to reach the success condition, while Poor is very likely to reach the fail condition.

There are two steps for the model creation:

  1. Complete the correlation analysis

  2. Choose factors and build the model

Correlation Analysis

The first step in creating a new predictive model is to conduct a correlation analysis. The correlation analysis is an automatic process done by Forwrd using sophisticated mathematical and statistical algorithms to determine the correlation of each factor (field) from our model base to the business objectives defined by the metrics.

To conduct the correlation analysis, follow the steps below:

  1. Go to projects and choose the project folder (or create a new one) you wish your model to be assigned to

  2. Click on 'New' to start a new correlation analysis and name your model

  3. Choose the base relevant for your model

  4. Choose the metrics relevant for your model

  5. Click on 'Create' and wait for the correlation analysis to complete (this process might take a few minutes)

  6. Once completed, you can see the results of the analysis:

    1. The list on the left side shows the factors (fields) that correlate with the desired outcome.

    2. When you select a factor, you can see how each value of this factor affects your metric. In the image below, you see that the Demo Request and Contact Us forms have a positive effect while the others have a negative effect.

    3. Note that the analysis picks the most relevant factors. You can manually add additional factors by clicking on 'Edit->Add' and selecting the ones you want to add.

Model Creation

  1. If you are ready to build your model, click on 'Build Model'. Choose which factors will be part of your model by checking or unchecking them, and click 'Build' when you are done. The platform will build the model based on your configurations (this process might take a few minutes).

  2. Your model is ready and you can review the model's prediction on your data.

  3. You can hover over each prediction and see the a clear explanation of why Forwrd decided to score it this way.

Did this answer your question?