Model Selection Guide
Picking the right model for a given problem
statement is an art as well as a science. In this
guide, we will provide some guidelines that
can help you select the right model for a
specific situation
We will use the following 4 criteria for model selection
Columns
Problem Type
Relationship
Regression Linear
Classification Non-linear
Data Type Data Size
Text Small
Numeric Large
Regression
Simple Linear Moderately Linear Highly Non-Linear
Relationship Relationship Relationship
Linear Polynomial Decision Tree Regressor
Regression Regression
Random Forest
Regressor
XGBoost Regressor
SVM Regressor
Regression
Small to Large
Medium Dataset Dataset
Linear Regression Random Forest
Regression
Polynomial Regression
XGBoost Regression
Decision Tree Regression
SVM Regressor
Classification
Linear Complex, Non-Linear
Relationship Relationship
Logistic Regression Decision Trees
Random Forest
XGBoost
Classification
Text Data Categorical Data Numeric Data
Naïve Bayes Decision Trees Logistic Regression
SVM Random Forest K-Nearest Neighbours
XGBoost SVM
Random Forest
XGBoost
Classification
Small to Large
Medium Dataset Dataset
Logistic Regression Random Forest
SVM XGBoost
Naïve Bayes
Decision Tree