This site was created to document Ken Michalek’s final project in MATH-E23C, a course offered at the Harvard University Extension School.

math23c_bot()

Here is the “documentation” for math23c_bot() and all its sub-functions. math23c_bot() is an R function that can take in a dataset as a parameter and employ several statistical methods (barplots, scatterplots, permutation tests, chi-square tests, etc.) to find relationships within the data. It can also utilize machine learning techniques (decision trees and neural nets) to make predictions.

Firefighting AI

Using the forestfires.csv dataset from the UCI Machine Learning Repository, the math23c_bot() function was able to find relationships between the time of week/year and the spread rate of forest fires in the Montesinho Natural Park, as well as use information on the temperature, rain, wind speed, and relative humidity to predict where a fire is most likely to occur in a 9x9 grid of the park on any given day.