Skip to content

This is a neural network that makes predictions about admissions selectivity based on universities' attributes.

Notifications You must be signed in to change notification settings

CarstenSavage47/admissions-predictions

Repository files navigation

Admissions Predictions

Overview

The purpose of this neural network is to predict whether a college is selective or not based on attributes.

The current model is using:

  • 75th percentile ACT scores
  • Admittance rate percentages
  • Total enrollment
  • Total out-of-state price
  • Percent of enrollment that is non-white
  • Historically black university status (dummy variable)
  • Percent of total enrollment that is women

These are useful predictors for whether a college is selective. I define a 'selective' university as one that has an acceptance rate of less than 50%.

Technical Notes

Evaluating kurtosis:

  • Mesokurtic: Data follows a normal distribution

  • Leptokurtic: Heavy tails on either side, indicating large outliers. Looks like Top-Thrill Dragster.

  • Playtkurtic: Flat tails indicate that there aren't many outliers.

  • A kurtosis value greater than +1 indicates the graph is very peaked. Leptokurtic.

  • A kurtosis value less than -1 indicates the graph is relatively flat. Playtkurtic.

  • A kurtosis value of 0 indicates that the graph follows a normal distribution. Mesokurtic.

Evaluating skewness:

  • A negative value indicates the tail is on the left side of the distribution.
  • A positive value indicates the tail is on the right side of the distribution.
  • A value of zero indicates that there is no skewness in the distribution; it's perfectly symmetrical.

Thank you Venelin (https://curiousily.com/posts/build-your-first-neural-network-with-pytorch/) and StatQuest (https://www.youtube.com/watch?v=FHdlXe1bSe4) for creating fantastic guides to PyTorch.

About

This is a neural network that makes predictions about admissions selectivity based on universities' attributes.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages