A two-part JMP Statistical Software Workshop co-sponsored by the Biostatistics Core Facility in the College of Health Sciences and the Department of Applied Economics and Statistics will be offered on Wednesday, Nov. 1 from 10 a.m. to 3 p.m.
The morning session will take place in the Atrium of the STAR Health Sciences Complex; the afternoon session will be directly across South College Avenue in 132 Townsend Hall. Faculty and graduate students with general interest in JMP capabilities or specific interest in enhancing their statistical skills are invited to register.
Topics covered will range from very basic descriptive summaries to advanced analyses. JMP is a SAS product characterized as a family of statistical discovery tools tailored to meet specific analysis needs. It is visually based, interactive, comprehensive and extensible.
It can be used as a front or back end with R and Matlab, and JMP can create and run SAS code if a SAS connection is set up. JMP is available for free, courtesy of the Department of Applied Economics and Statistics, to members of the University Community through UD Deploy and runs on both Macs and PCs.
Since its introduction on UD Deploy in 2012-2013, campus downloads have increased each year, with well over 4,000 in 2017-2018. Ease of use and powerful analytics make it a good choice for teaching and research.
Attend either one or both sessions as you like. Pizza provided for lunch. The sessions will continue in the afternoon in the College of Agriculture and Natural Resources. We encourage you to download JMP PRO before the workshop if you wish to follow along with the presentation, but this is optional. If you have no prior familiarity with JMP, we encourage you to view the Beginners Tutorial linked from the JMP PRO help menu. Registrants will be sent a link to access data featured during the workshop.
MORNING SESSION – STAR Atrium
10:00 Exploratory Data Analysis and Data Visualization
Featuring Graph Builder and Distribution platforms
Dynamically linked interactive graphics
Root cause analysis
Reporting and Sharing Results – exporting HTML5, Flash
Creating Web Reports and Dashboards
11:00 Decision Trees
Decision Tree – classification for categorical or regression for continuous responses
Bootstrap Forest – averaging many trees (JMP PRO)
Boosted Trees – combining a sequence of small tree (JMP PRO)
11:30 Text Exploration
Visualizing Text Data
Working with the Document Term Matrix
Creating and Modeling with Document and Topic Vectors (JMP PRO)
12:00 LUNCH – free pizza
AFTERNOON SESSION – 132 Townsend
1:00 JMP as Your Data Hub
Easily get data into JMP – from Excel, Text, Internet & Data Bases without coding
Cleanup messy and missing data – recode, outlier detection, data imputation
Connect JMP with SAS, R, MATLAB
Publish models in Python, C, Java Script, SQL or SAS (JMP PRO)
1:30 Building Better Models Using Robust Data Mining Methods to Prevent Overfitting
Honest Assessment Method – Train, Validation (Tune), and Test Data subsets
For smaller data sets use K-fold cross-validation or Penalization criteria
2:00 Regression Methods
Ordinary Least Squares, Stepwise Regression, Logistic Regression, General Linear Models
Generalized (Penalized) Regression (JMP PRO)
2:30 Neural Networks
Single layer of nodes with sigmoidal HTanh activation function
More flexible dual-layer of nodes plus Linear and Gaussian activation functions (JMP PRO)
Boosted neural network
For more information, visit the UD events website.