Enacting Recommendation

The Enacting Recommendation feature allows you to put Sanatio's recommendation into practice by creating a new model with the recommended changes.

A comparison markdown report is generated which indicates the improvements in the performance of the model before and after enacting the recommendations and stores the new model in the object's model attribute.

How to access the feature?

from sanatio.validations.sanatio_enact import enact

enact_object = enact(...parameters...)

Parameters required for initialization

Parameter name	Data Type	About
X_train	Data Frame	The training data used for the initial model trained
y_train	Data Series	The truth value used to train the initial model
X_test	Data Frame	The testing feature data split
y_test	Data Series	The ground truth data for testing
model	object	The initial model before enacting the recommendations
pearson_threshold	int - Optional	The pearson threshold filter you want for the future model
vif_factor	int - Optional	The vif threshold filter you want for the future model

Enact types

The different enact types implements recommendation as per the different routines validation type we have in Sanatio.

The different types of enact that are available in Sanatio with respective function are

• Logistic enact

  • logistic_enact()

• Linear enact

  • linear_enact()

• Tree classification enact

  • tree_classification_enact()

How to enact recommendation?

from sanatio.validations.sanatio_enact import enact

enact_object = enact(...parameters...)
enact_object.logistic_enact()

Example

from sklearn.datasets import load_breast_cancer
from sklearn.model_selection import train_test_split
import pandas as pd
from sklearn.linear_model import LogisticRegression

data = load_breast_cancer()
X,y = load_breast_cancer(return_X_y=True)
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.33, random_state=42)

lr =LogisticRegression()
lr.fit(X_train,y_train)
prediction = lr.predict(X_test)
one_class = lr.predict_proba(X_test)[:,1]
two_class = lr.predict_proba(X_test)

from sanatio.validations.sanatio_enact import enact

enact_obj = enact(X_train= pd.DataFrame(X_train,columns=data.feature_names),
                  y_train = y_train,
                  X_test = pd.DataFrame(X_test,columns=data.feature_names),
                  y_test = y_test,
                  model = lr, pearson_threshold = 0.9, vif_factor = 500
                 )
enact_obj.logistic_enact()