lift-提升度 计算

lift-提升度

• Lift衡量的是，与不利用模型相比，模型的预测能力“变好”了多少，结合二分类模型的混淆矩阵

  	预测positive	预测negative
  实际positive	TP	FN
  实际negative	FP	TN

  $$lift=\frac{precision}{实际正样本率}=\frac{\frac{TP}{TP+FP}}{\frac{TP+FN}{TP+TN+FN+FP}}$$

• 例子

  import pandas as pd
  import numpy as np
  from sklearn.metrics import precision_score as psy_true = np.array([1,1,0,0,0,0,0,0,0,0])
  y_pred = np.array([1,1,0,0,0,1,1,0,0,0])# 精准、查准、命中
  precision = ps(y_true,y_pred) #0.5
  # 真实正样本
  pos_rate = y_true.sum()/y_true.shape[0] # 0.2# 提升度
  lift = precision / pos_rate
  lift"""输出"""
  2.5
  

---