Boost Your Data Expertise - What's New in Minitab 19.2020.1

© 2020 Minitab, LLC.
2
Minitab Statistical Software
delivers business and predictive
analytics, paired with essential
visualizations to enable data-
driven value creating decision
making.
Everyone in an organization,
regardless of statistical
background, is empowered when
they have access to easy to use
analysis tools.

Predictive Analytics
3
encompasses a variety
of statistical techniques from data
mining, predictive modelling
and machine learning, that analyze current
and historical facts to
make predictions about future or otherwise
unknown events

Regression Challenges
Regression and Logistic Regression often don’t work well with larger
observational data sets.
► Everything is significant.
► Regression assumptions not met.
► Relationships are nonlinear.
► Extreme outliers.
► Many missing values.
4

5
The latest release of Minitab
Statistical Software expands upon
our Predictive Analytics offerings
✓ Classification and Regression Trees
✓ Model Validation
✓ Python Integration

Minitab Machine Learning Methods
6
Supervised learning in Minitab
Classification Algorithms
• Linear Discriminant Analysis (LDA)
• Quadratic Discriminant Analysis (QDA)
• Logistic Regression
• CART Classification Tree
Regression Algorithms
▪ Simple Regression
▪ Polynomial Regression
▪ Multiple Regression
▪ Nonlinear Regression
▪ Partial Least Squares
▪ CART Regression Trees
Unsupervised Learning in
Minitab
Clustering Algorithms
▪ Cluster Observations
▪ Cluster Variables
▪ Cluster K-Means
Data Reduction
Algorithms
▪ Principal Components
Analysis
▪ Factor Analysis

New Menu!! Classification and Regression Trees
7

CART (Classification and Regression Trees)
► Tree-based machine learning
► Builds classification and regression models in the
form of a tree structure
► Available for continuous, binary or multinomial
targets
► Ideal for large observational datasets
► Offers automatic variable selection
► Output is easy to interpret
8

Decision Trees
► Two types:
▪ Classification Trees: the
response is Binary or Multinomial
▪ Regression Trees: the response
is Continuous
► Constructed using recursive
partitioning
9

Terminology
► Root Node: The whole training
set. This is the start point of the
decision tree.
10

Terminology
► Splitter: the decision rule
that defines how the node is
partitioned
11
Splitter

Terminology
► Decision Node: Nodes that
split into sub-nodes.
12
Decision Node

Terminology
► Terminal Node: Nodes
that do not split. Also
called a Leaf.
13

Classification Tree Example: What’s contributing to out of spec product?
An engineer would like to understand what factors and settings contribute to product that
meets specification.

Response: Target

Results
16

Assess Model Accuracy
17

Relative Variable Importance
18

Receiver Operator Characteristic (ROC) Curve
► ROC evaluates event separation
when observations are ordered
based on the predicted
probabilities.
19

Flexibility to Easily Fit Alternative Trees
20

Regression Trees: Decision Trees with a Continuous Response
An analyst would like to understand the drivers of a high Net Promoter Score, which is a
measure of customer loyalty.
NPS: On a scale of 0 to 10, how likely is it that you would recommend our organization to a
friend or colleague?

Is it better to use Linear models or Tree-Based models?
► If the relationship between your x’s and y variables is well approximated by a linear model,
linear regression will outperform the tree-based model.
► If there is high non-linearity between your x and y variables, a tree model will outperform a
classical regression model.
► If your goal is to build a model that is easy to explain, a decision tree model will almost always
be a better choice.
24

© 2019 Minitab, Inc.
MODEL VALIDATION
Ensuring your models fit well
25

Validation
Data is usually split into a training set and a test set:
► Training (or learn) data is used to create the model and
estimate the model coefficients.
► Test data is used to assess model performance. It is used to
produce fitted values, which are then compared to the
corresponding observed responses in the test data to assess the
model’s predictive ability.
26

Validation: Reinforcing Predictions
► Assessing the “goodness” of a model with the same data that is used to fit the model leads to
overfitting and overfit models do not predict well.
► Validation is an honest assessment of how well the model will fit the new data.
27

Validation is easy to incorporate into your predictive analytics
28

Open-Source Integration
29
▪ Python integration
▪ Allow people to execute external Python scripts that use optional Minitab variables
(columns, constants, matrices) as inputs.
▪ Results are returned to Minitab and displayed in the Output Navigator and Output Pane.

Benefit: Enables Minitab to part of an organization’s analytics
community
► Python scripts run outside of Minitab
► The data can be in an external file or in the Minitab worksheet– this offers flexibility depending
on the type of data
▪ Python scripts can also use matrices or constants that are part of the worksheet
► Python connectivity will help easily extend the functionality within the software beyond what’s
possible in the interface or with a macro
30

Multi-Vari Chart
32

Variability Chart
33

Performance Improvements!
► Insert Rows
► Merge
► Move Columns
► Opening Large Excel files
► Tables
► Worksheet Stacking
► Boxplots with large amounts of data
► Graphical Summary with large amounts of data
34

User Interface Improvements
► Add maximize button to Graph Editing Window
35

Solutions Analytics
Solutions Analytics is our integrated approach to providing software and
services that enable organizations to make better decisions that drive
business excellence

.2020.1
► New Predictive Analytics Menu: Classification and Regression Trees
► Cross Validation for Regression and Binary Logistic Regression (BLR- Validation with a Test
set only)
► Multi-Vari chart and Variability Chart
► User Interface improvements
► Performance Improvements
► Python Integration
► Send to Minitab Workspace
► Additional Tools menu provides easy access to Quality Trainer and Minitab Workspace
37

Our Approach: More Than Business Analytics…Solutions Analytics
Software
Services
Training
Learn first-hand by attending public or
customized trainings in your facilities
according to your requirements.
Statistical
Consulting
Personalized help with statistical
challenges from collecting the right data
to interpreting analysis more.
Support
Assistance with installation,
implementation, version updates and
license management.
Master statistics and
Minitab anywhere
with online training
Machine learning and
predictive analytics
software
Start, track, manage
and execute
improvement projects
with real-time
dashboards
Powerful statistical
software everyone
can use.
Data Analysis Predictive Modeling Visual Business Tools Project Oversight
3
9
Visual tools to
process and product
excellence.
Online Training
Solutions analytics is our integrated approach to providing software and services that enable organizations to make better decisions that drive business excellence.

Boost Your Data Expertise - What's New in Minitab 19.2020.1

More Related Content

What's hot (15)

Similar to Boost Your Data Expertise - What's New in Minitab 19.2020.1 (20)

More from Minitab, LLC (16)

Recently uploaded (20)

Boost Your Data Expertise - What's New in Minitab 19.2020.1