See how to use JMP to analyze visually examine system failure data and build, compare and select models to predict system failures.
Presenter: Ghazal Falahi
See how to use JMP to analyze visually examine system failure data and build, compare and select models to predict system failures.
Presenter: Ghazal Falahi
The presenter uses public utility company transformer data to show how to visually explore the data and identify factors that may be main causes of failures. She uses and interprets Distribution, Graph Builder, Scatterplot Matrix, Pairwise Correlations, Color Map for Correlations and more.
The presenter gives and overview of the requirements for predictive modeling of complex data and uses JMP Pro to analyze and predict utility company transformer failures. She demonstrates how JMP Pro easily segments data into test, traiining and validation groups to avoid overfitting. She shows how to implement and interpret results from Decision Trees, Boosted Trees, Bootstrap Forests and Neural Networks. She concludes by demonstrating the use and value of Model Comparison and Contour Profile and how to use the Prediction Profiler and Simulators to examine what-if scenarios.