SHAP解释器SE共4篇
排序
SHAP解释器SE：Error in permshap.default(model rr, X= train set, feature names = xvars, : Permutation SHAP only supported for up to 14 features
报错原因:该模型最多支持14个特征进行shap计算解决方法：压缩特征数量
23天前
=", ">"), ret.second(split_index), :Unknown decision_type-决策链社区论坛" class="lazyload fit-cover radius8">Error in ifelse(decision_type %in% c(“>=”, “>”), ret.second(split_index), :Unknown decision_type
【报错原因】 对于lightgbm 数据存在数值类别，变成字符类别 【解决方法】 https://github.com/ModelOriented/treeshap/issues/28
4个月前
Error: You should have at least two distinct break values. Value cannot be null. (Parameter ‘s’)
【报错原因】模型拟合问题【解决方法】更换模型或变量
4个月前
Error in gbm.unify(model, data):Models built on data with categorical features are not supported – please encode them before training.
【报错原因】 类别数据需要encode 【解决方法】 前面链接 独热编码+标签转换编码+转换变量类型 进行encode
4个月前
