I use the following way in scikit-learn. Easy and straightforward.

scores= cross_val_score(clf, X, y, cv=10, scoring='accuracy')

avg_accuracy = scores.mean() #average accuracy

sd_accuracy = scores.std() #standard deviation

You can change the scoring to "precision_weighted" for obtaining precision scores of each fold and "recall_weighted" for recall scores of each fold. Why weighted? Read here more about the average parameter regarding these evaluation metrics: https://scikit-learn.org/stable/modules/generated/sklearn.metrics.precision_score.html

Hope that helps!

Standard deviation of 10 accuracy (for 10 fold) that you got is one best measure to validate your model. The standard deviation among accuracy must be less, which means that your model is reliably producing same predictive performance for any data segments.

Hi everyone.

Thank you Rabindra Lamsal.

Till now, I am using the following methods.

from sklearn.model_selection import cross_val_predict

from sklearn import metrics

y_pred = cross_val_predict(modelN,MyX,MyY,cv=10)

totacu=round((metrics.accuracy_score(MyY,y_pred)*100),3)

totMisacu=round((1-metrics.accuracy_score(MyY,y_pred))*100,3)

sensitivityVal=round((metrics.recall_score(MyY,y_pred))*100,3)

precision=round((metrics.precision_score(MyY,y_pred))*100,3);

f1score=round(2*((sensitivityVal*precision)/(sensitivityVal+precision)),3)

Just now I checked with the syntax what you specified.

from sklearn.model_selection import cross_val_score

Tscores= cross_val_score(modelN, MyX, MyY, cv=10, scoring='accuracy')

totacu = round((Tscores.mean()*100),3) #average accuracy

Rscores= cross_val_score(modelN, MyX, MyY, cv=10, scoring='recall_weighted')

sensitivityVal=round((Rscores.mean()*100),3)

Pscores= cross_val_score(modelN, MyX, MyY, cv=10, scoring='precision_weighted')

precision=round((Pscores.mean()*100),3)

f1score=round(2*((sensitivityVal*precision)/(sensitivityVal+precision)),3)

Both methods are giving the same results.

Everyone can use, any of one among these two..

thank you..