知识卡片
预处理拟合范围属于评价契约
若填补、缩放或分箱参数从全数据计算后才切分,留出信息可能影响训练。重构时先记录该行为,不能用高分合理化泄漏;若修正应作为有意行为变化,采用训练拟合再用于留出数据并重新建基线。
这张卡片看不懂?
这张卡片已通过人工审查,不再接受失败反馈。
已有反馈 · 0 条
还没有人提交反馈。
在整书笔记中的位置
- · 重构与技术债
原书出处 · 1 处
- · 查看原书依据
passengers = impute_nans(passengers, categorical_columns=["Embarked"], continuous_columns=["Fare", "Age"]) passengers = add_derived_title(passengers) passengers = add_is_alone_column(passengers) passengers = add_categorical_columns(passengers) passengers = passengers.drop(["Parch", "SibSp", "Name", "Passengerid", "Ticket", "Cabin"], axis=1) y = passengers["Survived"] X = passengers.drop("Survived", axis=1) X_train, X_test, y_train, y_test = train_test_split(X, y) model = train_model(RandomForestClassifier, X_train, y_train, n_estimators=100) return model, X_test, y_test model, X_test, y_test = prepare_data_and_train_model() Step 1. Run the notebook or code and ensure it works as