Background:

• Fire-Rated Doors: a door that can withstand 90 minutes of fires on one side(temperatures up to 2000 °F), and remain cool to touch on the other side. The door must also be able to retain its strength/sturdiness after 90 minutes.
• https://www.youtube.com/watch?v=UrYjyDNtwOE [link to sample fire rated door testing]

Project Overview:

• Based on the input of specific raw materials, our team should create an algorithm to indicate the expected, corresponding values pertaining to the properties of the resulting product(i.e find the correlation between input parameters in order to find the optimal strength parameters)
• Once the algorithm has been created, our team will create an application/program that allows a user to input the data parameters(raw materials) and get feedback on the output of expected properties.

[GUI & Jupyter Notebook]

Project Idea:

General Overview:

• Cleanup Data
  • Understand process model
  • Consider importance of presented feature parameters
  • Decide what are the target variables
  • Reformat data into a more cohesive model for visual neatness/understanding
• Understanding the Data
  • Feature Importance
• Implementation of Multi-Target Prediction
  • Many inputs, with many outputs needed to be predicted
  • Model Used: Random Forest
  • Each model predicts one target:
    • density 
    • compression
    • flexural
    • compression
    • adhesion
    • hardness
    • moisture
• GUI & Jupyter Notebook
• Deliverable to Client:
  • folder with:
    • how to add more data to model
    • how to install and properly run the Jupyter Notebook file