Text Analytics for Predicting Crisis Events in Veterans

Student: Wylie Frydrychowicz
Mentor: Dr. Praveen Madiraju

Project Description:

Veterans can have a hard time returning to civilized life. They can end up in various crises such as arrest, hospitalization, relapse, or angry outbursts. We want to help the veterans by predicting if they will be in crisis.

Project Goal:

Using data we collect from mobile applications that we built and continue to build, we want to create a machine learning algorithm that uses NLP (Natural Language Processing) to predict crisis events in veterans.

Summary of Milestones:

• Read research articles about:
  • Text/group messaging systems
  • Natural language processing
  • Veteran-related crises
• Gather:
  • Survey data
  • Feedback data
  • Text/group messaging data
• Implement:
  • Text/group messaging system
  • Machine learning algorithm that uses NLP
• Write the research paper
• Make the poster
• Present the poster

Milestones:

Week	Description
Week 1: Orientation	Meet with REU students and mentors Learn about introductory data science topics Meet with my mentor to talk about the project
Week 2: Android/iOS programming	Continue programming feedback activity (Android) Continue programming serverside feedback activity (Php, mysql) Start programming feedback activity (iOS) TBD: Collect data from apps
Week 3: Group Messaging	Read articles about implementing a group messaging system Find best practices for implementing one TBD: Collect data from the system
Week 4: Natural Language Processing	Read articles about NLP Find best practices for implementing one Practice implementing NLP with a kaggle dataset
Week 5: TBD - Collect data	Hopefully have veterans use apps to get data Will update wiki when I know we'll have veterans
Week 6: TBD - Extract data	Implement PHP/python script to export data into Excel or Csv file
Week 7: NLP on Data	Using the data, create a ML alg that uses NLP Use survey data, feedback data, and group message data
Week 8: Research Paper	Plan to work on the paper throughout the whole summer, but this week will be used for writing the majority of my paper and refining it since I will have my data and ML alg output
Week 9: Poster	Create poster by taking info from paper
Week 10: Finale	Finish up critiquing paper/poster Present poster