SaniyaAbushakimova/Rigorous-Modeling-Techniques-for-Estimating-Student-Reaction-Times

Project completed on May 16, 2024.

Project description

Using Reaction Time Survey dataset conduct a rigorous regression modeling and analysis to estimate student reaction times. The project outline is as follows (more details in project_report.pdf):

1. Abstract
2. Exploratory Data Analysis (EDA)
   2.1. Data Understanding
   2.2. Data Insights
   2.3. Data Pre-processing
3. Model Building
   3.1. Variable Selection and Model Fitting
   3.2. Diagnostics and Remedies
   a) Unusual observations
   b) Error assumptions
   c) Structure assumptions
4. Model Comparison and Selection
   4.1. A model with an interaction term
   4.2. LASSO Regression
5. Discussion of Results and Conclusion
   5.1. Summary
   5.2. Challenges and Next Steps
   5.3. Reflection on Lessons Learned

Regression Analysis tools used in this project

• Adjusted R^2
• VIF
• Pearson correlation
• ANOVA
• Cramer's V association
• Forward variable selection
• Lasso Regression
• Diagnostics/Remedies
  • Mahalanobis Distance
  • Studentized Residual Test
  • Cook's Distance
  • Q-Q plot / Shapiro-Wilk Test
  • Residuals vs Fitted plot / Breusch-Pagan Test
  • Residuals vs Index plot / Durbin-Watson Test
  • Added-Variable plots
  • Box-Cox transformation

Other details

survey.csv -- raw dataset
survey_postEDA.csv -- dataset after cleaning and preprocessing