R语言基于glmnet构建分类模型并可视化特征系数（coefficient）以及L1正则化系数（lambda）实战

Posted 2021-09-08 Data+Science+Insight

R语言基于glmnet构建分类模型并可视化特征系数（coefficient）以及L1正则化系数（lambda）实战

# 导入测试数据集

data(BinomialExample)
x <- BinomialExample$x
y <- BinomialExample$y

# 构建模型并可视化系数

glmmod <- glmnet(x, y, alpha=1, family="binomial",lambda = logspace(-10, 1, 100))

# Plot variable coefficients vs. shrinkage parameter lambda.
plot(glmmod, xvar="lambda")
abline(v = cv.glmnet(x,y,lambda = logspace(-10, 1, 100))$lambda.min, col = "black")