机器学习中的工作流机制

在项目开发的时候，经常需要我们选择使用哪一种模型。同样的数据，可能决策树效果不错，朴素贝叶斯也不错，SVM也挺好。有没有一种方法能够让我们用一份数据，同时训练多个模型，并用某种直观的方式（包括模型得分），观察到模型在既有数据上的效果？有的，管线工作流pipeline就是专门干这个的，再配上决策边界，所有模型只用一眼，就能确定优劣，选择你的梦中情模。上效果图。

在这里插入图片描述

分为两行，上面是sklearn自带数据集中的数据，分两类。从第二列开始，每一列是某种模型在当前数据集中的拟合效果。如何查看某种模型效果好坏？从两个方面，左上角的模型得分，和图中颜色深浅，两种颜色的分解代表模型的决策边界。

下面是笔者自己的数据，分为4类。同样不同颜色的分界代表两种类型的判别边界。如果只看模型得分，那得分为100%的模型有5个，选再根据决策边界进一步确定更优秀的模型，为工程所用。这里贴出笔者所用代码供各位修改，也可以直接取官方代码修改

def loadTrainData():df = pd.read_csv('./your/dataset/path/data.csv')trainDataLabel = df.valuesnodeData = trainDataLabel[:, :2], trainDataLabel[:, -1]return nodeDatadef trainAnalySave():from matplotlib.colors import ListedColormapimport joblibfrom sklearn.datasets import make_circles, make_classification, make_moonsfrom sklearn.discriminant_analysis import QuadraticDiscriminantAnalysisfrom sklearn.ensemble import AdaBoostClassifier, RandomForestClassifierfrom sklearn.gaussian_process import GaussianProcessClassifierfrom sklearn.gaussian_process.kernels import RBFfrom sklearn.inspection import DecisionBoundaryDisplayfrom sklearn.model_selection import train_test_splitfrom sklearn.naive_bayes import GaussianNBfrom sklearn.neighbors import KNeighborsClassifierfrom sklearn.neural_network import MLPClassifierfrom sklearn.pipeline import make_pipelinefrom sklearn.preprocessing import StandardScalerfrom sklearn.svm import SVCfrom sklearn.tree import DecisionTreeClassifiernames = ["Nearest Neighbors","Linear SVM","RBF SVM","Gaussian Process","Decision Tree","Random Forest","Neural Net","AdaBoost","Naive Bayes","QDA",]classifiers = [KNeighborsClassifier(3),SVC(kernel="linear", C=0.025),SVC(gamma=2, C=1),GaussianProcessClassifier(1.0 * RBF(1.0)),DecisionTreeClassifier(max_depth=5),RandomForestClassifier(max_depth=5, n_estimators=10, max_features=1),MLPClassifier(alpha=1, max_iter=1000),AdaBoostClassifier(),GaussianNB(),QuadraticDiscriminantAnalysis(),]# X, y = make_classification(#     n_features=2, n_redundant=0, n_informative=2, random_state=1, n_clusters_per_class=1# )# rng = np.random.RandomState(2)# X += 2 * rng.uniform(size=X.shape)# linearly_separable = (X, y)nodeData = loadTrainData()datasets = [# make_moons(noise=0.3, random_state=0),make_circles(noise=0.2, factor=0.5, random_state=1),# linearly_separable,nodeData,]# figure = plt.figure(figsize=(27, 9))figure = plt.figure(figsize=(15, 4))i = 1# iterate over datasetsfor ds_cnt, ds in enumerate(datasets):# preprocess dataset, split into training and test partX, y = dsX_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.4, random_state=42)x_min, x_max = X[:, 0].min() - 0.5, X[:, 0].max() + 0.5y_min, y_max = X[:, 1].min() - 0.5, X[:, 1].max() + 0.5# just plot the dataset firstcm = plt.cm.RdBucm_bright = ListedColormap(["#FF0000", "#00FF00", "#FFFF00", "#0000FF"])ax = plt.subplot(len(datasets), len(classifiers) + 1, i)if ds_cnt == 0:ax.set_title("Input data")# Plot the training pointsax.scatter(X_train[:, 0], X_train[:, 1], c=y_train, cmap=cm_bright, edgecolors="k")# Plot the testing pointsax.scatter(X_test[:, 0], X_test[:, 1], c=y_test, cmap=cm_bright, alpha=0.6, edgecolors="k")ax.set_xlim(x_min, x_max)ax.set_ylim(y_min, y_max)ax.set_xticks(())ax.set_yticks(())i += 1# iterate over classifiersfor name, clf in zip(names, classifiers):ax = plt.subplot(len(datasets), len(classifiers) + 1, i)clf = make_pipeline(StandardScaler(), clf)clf.fit(X_train, y_train)score = clf.score(X_test, y_test)# DecisionBoundaryDisplay.from_estimator(#     clf, X, cmap=cm, alpha=0.8, ax=ax, eps=0.5# )# save satisfied modelsavedPath = r'..\models\sklearn\\'savedList = ["Nearest Neighbors", "RBF SVM", "Neural Net"]if name in savedList:joblib.dump(clf, savedPath + name + '.pkl')# Plot the training pointsax.scatter(X_train[:, 0], X_train[:, 1], c=y_train, cmap=cm_bright, edgecolors="k")# Plot the testing pointsax.scatter(X_test[:, 0],X_test[:, 1],c=y_test,cmap=cm_bright,edgecolors="k",alpha=0.6,)ax.set_xlim(x_min, x_max)ax.set_ylim(y_min, y_max)ax.set_xticks(())ax.set_yticks(())if ds_cnt == 0:ax.set_title(name)ax.text(# x_max - 0.3,# y_min + 0.3,x_min + 0.4,y_max - 0.4 - ds_cnt,("%.2f" % score),# ("%.2f" % score).lstrip("0"),# size=15,size=10,# horizontalalignment="right",horizontalalignment="left",)i += 1plt.tight_layout()plt.show()   nodeData = loadTrainData()
if __name__ == '__main__':trainAnalySave()