Scatter

class mastml.plots.Scatter

Bases: object

Class to generate scatter plots, such as parity plots showing true vs. predicted data values

Args:

None

Methods:

plot_predicted_vs_true: method to plot a parity plot

Args:

y_true: (pd.Series), series of true y data

y_pred: (pd.Series), series of predicted y data

savepath: (str), string denoting the save path for the figure image

data_type: (str), string denoting the data type (e.g. train, test, leaveout)

x_label: (str), string denoting the true and predicted property name

metrics_list: (list), list of strings of metric names to evaluate and include on the figure

show_figure: (bool), whether or not to show the figure output (e.g. when using Jupyter notebook)

Returns:

None

plot_best_worst_split: method to find the best and worst split in an evaluation set and plot them together

Args:

savepath: (str), string denoting the save path for the figure image

data_type: (str), string denoting the data type (e.g. train, test, leaveout)

x_label: (str), string denoting the true and predicted property name

metrics_list: (list), list of strings of metric names to evaluate and include on the figure

show_figure: (bool), whether or not to show the figure output (e.g. when using Jupyter notebook)

Returns:

None

plot_best_worst_per_point: method to find all of the best and worst data points from an evaluation set and plot them together

Args:

savepath: (str), string denoting the save path for the figure image

data_type: (str), string denoting the data type (e.g. train, test, leaveout)

x_label: (str), string denoting the true and predicted property name

metrics_list: (list), list of strings of metric names to evaluate and include on the figure

show_figure: (bool), whether or not to show the figure output (e.g. when using Jupyter notebook)

Returns:

None

plot_predicted_vs_true_bars: method to plot the average predicted value of each data point from an evaluation set with error bars denoting the standard deviation in predicted values

Args:

savepath: (str), string denoting the save path for the figure image

data_type: (str), string denoting the data type (e.g. train, test, leaveout)

x_label: (str), string denoting the true and predicted property name

metrics_list: (list), list of strings of metric names to evaluate and include on the figure

show_figure: (bool), whether or not to show the figure output (e.g. when using Jupyter notebook)

Returns:

None

plot_metric_vs_group: method to plot the metric value for each group during e.g. a LeaveOneGroupOut data split

Args:

savepath: (str), string denoting the save path for the figure image

data_type: (str), string denoting the data type (e.g. train, test, leaveout)

show_figure: (bool), whether or not to show the figure output (e.g. when using Jupyter notebook)

Returns:

None

Methods Summary

plot_best_worst_per_point(savepath, ...[, ...])
plot_best_worst_split(savepath, data_type, ...)
plot_metric_vs_group(groups, stats_group_df, ...)
plot_predicted_vs_true(y_true, y_pred, ...)
plot_predicted_vs_true_bars(savepath, ...[, ...])

Methods Documentation

classmethod plot_best_worst_per_point(savepath, data_type, x_label, metrics_list, show_figure=False, file_extension='.csv', image_dpi=250)

classmethod plot_best_worst_split(savepath, data_type, x_label, metrics_list, show_figure=False, file_extension='.csv', image_dpi=250)

classmethod plot_metric_vs_group(groups, stats_group_df, metrics_list, savepath, data_type, show_figure, file_extension='.csv', image_dpi=250)

classmethod plot_predicted_vs_true(y_true, y_pred, savepath, data_type, x_label, metrics_list=None, show_figure=False, ebars=None, file_extension='.csv', image_dpi=250, groups=None)

classmethod plot_predicted_vs_true_bars(savepath, x_label, data_type, metrics_list, show_figure=False, ebars=None, file_extension='.csv', image_dpi=250, groups=None)