model_inspector's methods for plotting predictions for models with two features involve making predictions for a grid of possible feature values. Those grids are numpy arrays, so they do not have feature names. When the model is trained using a pandas DataFrame, the result is that those plotting methods give ugly warnings that are particularly distracting for beginners:
I think the best solution is to turn the grids into DataFrames with column names from X.
model_inspector
's methods for plotting predictions for models with two features involve making predictions for a grid of possible feature values. Those grids are numpy arrays, so they do not have feature names. When the model is trained using a pandas DataFrame, the result is that those plotting methods give ugly warnings that are particularly distracting for beginners:I think the best solution is to turn the grids into DataFrames with column names from
X
.