17.8.1 Predictive Modeling Advisor Dialog Box

Contents

Introduction

Predictive Modeling Advisor is a build-in app in Origin, It helps you to pick the most appropriate predictive modeling that is best suited for your data, including the tools and Apps. If you are all new to Origin or are unfamiliar with the predictive modeling, this app could help you get started. Open this dialog, you could see three column, including What do you want to do, Your options and Suggested feature(s).

Performing Predictive Modeling Advisor

To perform Predictive Modeling Advisor tool from the menu:


There are three parts in this dialog:

  1. What do you want to do?
    Click to select the most appropriate answer, then a new question (if there is one) will appear in the second column Your options.
  2. Your options
    Click to select your answers to the successive questions until the suggested feature(s) appear in the third column.
  3. Suggested Feature(s)
    • Based on your answers, this App will list all related features in this column.
    • The graph icon and the App icon followed the option could help to distinguish the types of the feature. Also, you could tell if the feature is only available in Origin Pro by the Pro icon.
    • Click to select one of the features, then a note of the feature will appear at the bottom of this column.
    • Click Help button to open the corresponding help page in the browser.
    • Click Open Dialog button to open the dialog of the feature you select.
    • If you select an App but you still haven't installed it, Open Dialog button will turn into Install App button. Click Install App button to install the App directly.

Suggested feature(s)

What do you want to do? You options Suggested feature(s)
Classification Tree-base Models
Linear Models
Kernel Methods
Regressing Linear Models
Generalized Models
Mixed / Hierarchical Models
Robust & Distribution-aware Models
Nonlinear & Flexible Models
Tree-base Models
Clustering
Dimensionality Reduction Variance-based
Latent Variable Models
Independent/ Canonical Methods
Distance / Ordination-based