Concrete Strength prediction

Check out the script for the models that are created to predict the strength of concrete. It includes every step that is required to be followed while creating a machine learning model.

Uses:

1)Predict the strength of concrete with 94% accuracy.

2)Understand which independent factors have major impact on the strength of concrete

3)Strength prediction of concrete would help to get an idea about the age of construction projetcs where conceret is used as a material.

Procedure:

1) Data Preprocessing, EDA and Data Engineering

Libraries used: Numpy, Pandas, Matplotlib, Scipy, autoviz

2) Model Building

Libraries used: Scikit-learn, Statsmodels

3)Challenges (Multicollinearity & Heteroscedasticity)

Libraries used: Statsmodels, Scikit-learn

4) Feature selection and Overfitting

Libraries used: Scikit-learn, mlxtend, Statsmodels

5) Determine which model is the best

Libraries used: Pycaret, Xgboost

Results:

1) Training accuracy is 97% i.e coefficient of determination.

2) Testing accuracy is 94% i.e coefficient of determination.