/gridSearch-parameter-tuning-on-various-classification-model

using Grid search algorithm with cross validation to tune parameters and training for various classification models and compare between them

Primary LanguageJupyter Notebook

gridSearch-parameter-tuning-on-various-classification-model

using Grid search algorithm with cross validation to tune parameters and training for various classification models and compare between them this research aplied on digits by sklearn datasets

used models with tuned parameters

1-svm (C,kernel) 2-random forest (n_estimators) 3-logistic regression (C) 4-naive_bayes.GaussianNB (var_smoothing) 5-naive_bayes.MultinomialNB (alpha) 6-DecisionTree (criterion)

and used 5 folds for cross validation

result

1 svm scored:94.7697% with parameters values ={'C': 1, 'kernel': 'linear'}
2 random_forest scored:93.9362% with parameters values ={'n_estimators': 100}
3 logistic_regression scored:92.2114% with parameters values ={'C': 1}
4 nbm scored:87.0350% with parameters values ={'alpha': 0}
5 dt scored:81.0258% with parameters values ={'criterion': 'entropy'}
6 nbg scored:80.6928% with parameters values ={'var_smoothing': 1e-09}