sklearn.feature_selection讲解

本文详细介绍了sklearn中SelectKBest类的使用方法,该类用于根据特征评分选择最高得分的K个特征,适用于分类任务。文章通过实例展示了如何使用SelectKBest结合卡方检验对digits数据集进行特征选择。

摘要生成于 C知道 ,由 DeepSeek-R1 满血版支持, 前往体验 >

class sklearn.feature_selection.SelectKBest(score_func=, k=10)
作用:Select features according to the k highest scores
选出分数最高的k个特征

Parameters:
score_func : callable
Function taking two arrays X and y, and returning a pair of arrays (scores, pvalues) or a single array with scores. Default is f_classif (see below “See also”). The default function only works with classification tasks.
k : int or “all”, optional, default=10
Number of top features to select. The “all” option bypasses selection, for use in a parameter search.
输出分数最高的K个特征

类方法:
fit(X, y) Run score function on (X, y) and get the appropriate features.
对X,y数据的特征进行评价
fit_transform(X[, y]) Fit to data, then transform it.
只保留数据X的前K个分数最高的特征

examples:

>>> from sklearn.datasets import load_digits
>>> from sklearn.feature_selection import SelectKBest, chi2
>>> X, y = load_digits(return_X_y=True)
>>> X.shape
(1797, 64)
>>> X_new = SelectKBest(chi2, k=20).fit_transform(X, y)
>>> X_new.shape
(1797, 20)
请教学式按句详细讲解以下代码:###--------------------KNN算法与决策树算法-------------------- from sklearn.feature_extraction.text import TfidfVectorizer from sklearn.preprocessing import StandardScaler from sklearn.model_selection import train_test_split # 将文本数据转化为数值特征 vectorizer = TfidfVectorizer() X = vectorizer.fit_transform(data_str_list) # 划分数据集为训练集和测试集 X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42) # 特征缩放 scaler = StandardScaler() X_train = scaler.fit_transform(X_train.toarray()) X_test = scaler.transform(X_test.toarray()) from sklearn.neighbors import KNeighborsClassifier from sklearn.tree import DecisionTreeClassifier from sklearn.model_selection import GridSearchCV from sklearn.metrics import accuracy_score # 使用网格搜索进行超参数调优 param_grid = { "n_neighbors": [3, 5, 7, 9], "weights": ["uniform", "distance"], "algorithm": ["auto", "ball_tree", "kd_tree", "brute"] } knn = KNeighborsClassifier() grid_search = GridSearchCV(knn, param_grid, cv=5) grid_search.fit(X_train, y_train) print("KNN最优参数:", grid_search.best_params_) param_grid = { "criterion": ["gini", "entropy"], "max_depth": [3, 5, 7, 9] } dt = DecisionTreeClassifier() grid_search = GridSearchCV(dt, param_grid, cv=5) grid_search.fit(X_train, y_train) print("决策树最优参数:", grid_search.best_params_) # 训练分类器并进行预测 knn = KNeighborsClassifier(n_neighbors=5, weights="uniform", algorithm="auto") knn.fit(X_train, y_train) knn_pred = knn.predict(X_test) dt = DecisionTreeClassifier(criterion="gini", max_depth=9) dt.fit(X_train, y_train) dt_pred = dt.predict(X_test) # 混合使用KNN和决策树进行文本分类 ensemble_pred = [] for i in range(len(knn_pred)): if knn_pred[i] == dt_pred[i]: ensemble_pred.append(knn_pred[i]) else: ensemble_pred.append(knn_pred[i]) # 输出分类结果和准确率 print("KNN准确率:", accuracy_score(y_test, knn_pred)) print("决策树准确率:", accuracy_score(y_test, dt_pred)) print("混合使用准确率:", accuracy_score(y_test, ensemble_pred))
05-31
import pandas as pd import numpy as np import matplotlib.pyplot as plt import seaborn as sns from sklearn.ensemble import RandomForestClassifier from sklearn.model_selection import train_test_split, cross_val_score, GridSearchCV, learning_curve from sklearn.metrics import classification_report, confusion_matrix, accuracy_score, roc_curve, auc from sklearn.preprocessing import StandardScaler import joblib import os def load_and_preprocess_data(csv_path): """ 加载并预处理数据 Args: csv_path: CSV文件路径 Returns: tuple: (X_scaled, y, scaler, feature_names) """ print(f"Loading data from {csv_path}...") df = pd.read_csv(csv_path) # 打印数据集基本信息 print("\nDataset Overview:") print(f"Number of samples: {df.shape[0]}") print(f"Number of features: {df.shape[1]-2}") # 排除image_name和indicator print("\nFeature Statistics:") print(df[['num_spots', 'max_spot_area', 'total_spot_area', 'indicator']].describe()) # 检查缺失值 if df.isnull().sum().sum() > 0: print("\nWarning: Dataset contains missing values") print(df.isnull().sum()) # 处理缺失值 df = df.dropna() # 准备特征和目标 X = df[['num_spots', 'max_spot_area', 'total_spot_area']] y = df['indicator'] # 缩放特征 scaler = StandardScaler() X_scaled = scaler.fit_transform(X) return X_scaled, y, scaler, X.columns def evaluate_model(model, X_test, y_test, feature_names=None): """ 评估模型性能 Args: model: 训练好的模型 X_test: 测试特征 y_test: 测试标签 feature_names: 特征名称列表 Returns: dict: 包含评估结果的字典 """ y_pred = model.predict(X_test) print("\nModel Evaluation:") print(f"Accuracy: {accuracy_score(y_test, y_pred):.4f}") print("\nClassification Report:") print(classification_report(y_test, y_pred)) print("\nConfusion Matrix:") cm = confusion_matrix(y_test, y_pred) print(cm) results = { 'acc
03-23
"""用户价值分层预测全流程示例""" import pandas as pd import numpy as np from sklearn.preprocessing import OrdinalEncoder, StandardScaler from sklearn.impute import SimpleImputer from sklearn.model_selection import train_test_split from factor_analyzer import FactorAnalyzer from factor_analyzer.factor_analyzer import calculate_kmo from xgboost import XGBClassifier import lightgbm as lgb from sklearn.metrics import classification_report # ====================== # 1. 数据读取与清洗 # ====================== def load_and_clean(data_path): # 读取数据 df = pd.read_csv(data_path) # 缺失值处理 num_imputer = SimpleImputer(strategy='median') cat_imputer = SimpleImputer(strategy='most_frequent') # 数值型字段 numeric_cols = ['付费金额', '活跃时长', '广告收入', '留存'] df[numeric_cols] = num_imputer.fit_transform(df[numeric_cols]) # 分类型字段 categorical_cols = ['设备价值档位', '用户初始广告档位'] df[categorical_cols] = cat_imputer.fit_transform(df[categorical_cols]) # 异常值处理 df['活跃时长'] = np.where(df['活跃时长'] > 24, 24, df['活跃时长']) df['付费金额'] = np.where( df['付费金额'] > df['付费金额'].quantile(0.99), df['付费金额'].quantile(0.95), df['付费金额'] ) return df # ====================== # 2. 特征工程 # ====================== def feature_engineering(df): # 构造复合特征 df['ARPU密度'] = df['付费金额'] / (df['活跃天数'] + 1) df['广告展示率'] = df['广告曝光次数'] / df['短剧观看次数'] df['内容互动指数'] = np.log1p(df['收藏数'] * 1 + df['分享数'] * 2 + df['评论数'] * 3) # 分类变量编码 encoder = OrdinalEncoder(handle_unknown='use_encoded_value', unknown_value=-1) cat_cols = ['设备价值档位', '用户初始广告档位'] df[cat_cols] = encoder.fit_transform(df[cat_cols]) return df # ====================== # 3. 特征筛选与降维 # ====================== def feature_selection(df, target_col='付费意愿档位'): # 划分特征和目标 X = df.drop(columns=[target_col]) y = df[target_col] # 计算IV值筛选 from sklearn.feature_selection import mutual_info_classif iv_values = mutual_info_classif(X, y) iv_df = pd.DataFrame({'feature': X.columns, 'iv': iv_values}) selected_features = iv_df[iv_df['iv'] > 0.02]['feature'].tolist() X_selected = X[selected_features] # 因子分析降维 kmo_all, kmo_model = calculate_kmo(X_selected) if kmo_model > 0.6: fa = FactorAnalyzer(n_factors=5, rotation='varimax') fa.fit(X_selected) factor_scores = fa.transform(X_selected) factor_cols = [f'Factor_{i}' for i in range(1, 6)] X_factors = pd.DataFrame(factor_scores, columns=factor_cols) else: X_factors = X_selected.copy() return X_factors, y # ====================== # 4. XGBoost特征筛选 # ====================== def xgb_feature_importance(X, y): # 训练XGBoost模型 model = XGBClassifier( objective='multi:softmax', eval_metric='mlogloss', use_label_encoder=False ) model.fit(X, y) # 获取特征重要性 importance = pd.DataFrame({ 'feature': X.columns, 'importance': model.feature_importances_ }).sort_values('importance', ascending=False) top10_features = importance.head(10)['feature'].tolist() return X[top10_features] # ====================== # 5. LightGBM建模预测 # ====================== def lgb_modeling(X, y): # 数据划分 X_train, X_test, y_train, y_test = train_test_split( X, y, test_size=0.2, stratify=y, random_state=42 ) # 模型训练 model = lgb.LGBMClassifier( num_leaves=31, max_depth=5, learning_rate=0.1, n_estimators=300, class_weight='balanced' ) model.fit(X_train, y_train) # 模型评估 y_pred = model.predict(X_test) print(classification_report(y_test, y_pred)) return model # ====================== # 主流程执行 # ====================== if __name__ == "__main__": # 数据路径 DATA_PATH = "new.csv" # 执行流程 df = load_and_clean(DATA_PATH) df = feature_engineering(df) X, y = feature_selection(df) X_top10 = xgb_feature_importance(X, y) final_model = lgb_modeling(X_top10, y) # 模型保存 final_model.booster_.save_model('user_value_model.txt')详细讲解这段代码
最新发布
05-12
评论
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值