Machine Learning
Supervised Learning |
Unsupervised learning |
The model maps input to an output based on the previous input-output pairs |
No training is given to the model and it has to discover the features of input by self training mechanism. |
Scikit learn can be used in Classi¬fic¬ation, Regres¬sion, Cluste¬ring, Dimens¬ion¬ality reduct¬ion¬,Model Selection and prepro¬cessing by supervised and unsupe¬rvised training models.
Basic Commands
>>> from sklearn import neighbors, datasets, preprocessing |
>>> from sklearn.model_selection import train_test_split |
>>> from sklearn.metrics import accuracy_score |
>>> iris = datasets.load_iris() |
>>> X, y = iris.data[:, :2], iris.target |
>>> X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=33) |
>>> scaler = preprocessing.StandardScaler().fit(X_train) |
>>> X_train = scaler.transform(X_train) |
>>> X_test = scaler.transform(X_test) |
>>> knn = neighbors.KNeighborsClassifier(n_neighbors=5) |
>>> knn.fit(X_train, y_train) |
>>> y_pred = knn.predict(X_test) |
>>> accuracy_score(y_test, y_pred) |
Loading Data example
>>> import numpy as np |
>>> X = np.random.random((20,2)) |
>>> y = np.array(['A','B','C','D','E','F','G','A','C','A','B']) |
>>> X[X < 0.7] = 0 |
The data being loaded should be numeric and has to be stored as NumPy arrays or SciPy sparse matrices.
|
|
Processing Loaded Data
Standardization |
Normalization |
Binarization |
>>> from sklearn.preprocessing import StandardScaler |
>>> from sklearn.preprocessing import Normalizer |
>>> from sklearn.preprocessing import Binarizer |
>>> scaler = StandardScaler().fit(X_train) |
>>> scaler = Normalizer().fit(X_train) |
>>> binarizer = Binarizer(threshold=0.0).fit(X) |
>>> standardized_X = scaler.transform(X_train) |
>>> normalized_X = scaler.transform(X_train) |
>>> binary_X = binarizer.transform(X) |
>>> standardized_X_test = scaler.transform(X_test) |
>>> normalized_X_test = scaler.transform(X_test) |
Training And Test Data
>>> from sklearn.model_selection import train_test_split |
>>> X_train, X_test, y_train, y_test = train_test_split(X,y,random_state=0) |
Creating Model
Supervised Learning Estimators |
Linear Regression |
Support Vector Machines (SVM) |
Naive Bayes |
>>> from sklearn.linear_model import LinearRegression |
>>> from sklearn.svm import SVC |
>>> from sklearn.naive_bayes import GaussianNB |
>>> lr = LinearRegression(normalize=True) |
>>> svc = SVC(kernel='linear') |
>>> gnb = GaussianNB() |
Creating Model
Unsupervised Learning Estimators |
Principal Component Analysis (PCA) |
K Means |
>>> from sklearn.decomposition import PCA |
>>> from sklearn.cluster import KMeans |
>>> pca = PCA(n_components=0.95) |
>>> k_means = KMeans(n_clusters=3, random_state=0) |
|
|
Model Fitting
Supervised Learning |
Unsupervised learning |
>>> lr.fit(X, y) |
>>> k_means.fit(X_train) |
>>> knn.fit(X_train, y_train) |
>>> pca_model = pca.fit_transform(X_train) |
>>> svc.fit(X_train, y_train) |
Predicting output
Supervised Estimators |
Unsupervised Estimators |
>>> y_pred = svc.predict(np.random.random((2,5))) |
>>> y_pred = k_means.predict(X_test) |
>>> y_pred = lr.predict(X_test) |
>>> y_pred = knn.predict_proba(X_test)) |
Classification Metrics Model Performance
Accuracy Score |
Classification Report |
Confusion Matrix |
>>> knn.score(X_test, y_test) |
>>> from sklearn.metrics import classification_report |
>>> from sklearn.metrics import confusion_matrix |
>>> from sklearn.metrics import accuracy_score |
>>> print(classification_report(y_test, y_pred))) |
>>> print(confusion_matrix(y_test, y_pred))) |
>>> accuracy_score(y_test, y_pred) |
Clustering Metrics Model Performance
Adjusted Rand Index |
Homogeneity |
Cross-Validation |
>>> from sklearn.metrics import adjusted_rand_score |
>>> from sklearn.metrics import homogeneity_score |
>>> print(cross_val_score(knn, X_train, y_train, cv=4)) |
>>> adjusted_rand_score(y_true, y_pred)) |
>>> homogeneity_score(y_true, y_pred)) |
>>> print(cross_val_score(lr, X, y, cv=2)) |
|
