A data science project on predicting the compressive strength of concrete using different proportions of ingredients with regression models. Open the Technical_Report.ipynb to view the full report.
-
Concrete compressive strength is non-linear because linear models perform poorly when it is fitted.
-
For the full dataset, the best model is Random Forest with
$R$ 2 = 0.9205 and age as the top predictor. -
For the age >= 28, the best model is Random Forest with an
$R$ 2 = 0.9049 for the full dataset and water-to-cement ratio as the top predictor. -
A distinction was made between the inclusion of partially cured concrete due to different use cases.
-
Implementation uses GridSearchCV, RepeatedKFold and Pipeline from the scikit-learn library.
A partial fulfillment for the requirements for AIM MSDS 2022 Machine Learning 1 under Prof. Chris Monterola.