为什么同样的数据集同样的参数重复运行这行代码后，会得到两个准确率不一样的模型


from sklearn.ensemble import StackingClassifier
log_clf = LogisticRegression()
svm_clf = SVC(probability=True)
tree_clf = DecisionTreeClassifier()
log_ensemble = LogisticRegression()
stk_clf = StackingClassifier(
      estimators=[('lr', log_clf), ('svc', svm_clf), ('tree', tree_clf)],
    final_estimator=log_ensemble)
stk_clf.fit(train_features, train_target)
stk_clf.score(test_features, test_target)

为什么同样的数据集同样的参数重复运行这行代码后，会得到两个准确率不一样的模型？

  • 帮你找了个相似的问题, 你可以看下: https://ask.csdn.net/questions/7660085
  • 这篇博客也不错, 你可以看下【推荐系统】基于协同过滤的推荐系统示例
  • 同时，你还可以查看手册：在玩具数据集上比较不同的聚类算法 中的内容
  • 除此之外, 这篇博客: 商品销售额预测中的 获取不同颜色产品的平均指标 部分也许能够解决你的问题, 你可以仔细阅读以下内容或跳转源博客中阅读:
  • product_descriptor = ['product_type', 'product_gender', 'macro_function', 
                      'function', 'sub_function', 'model', 'aesthetic_sub_line', 'macro_material',
                      'month']
product_target_sum = train.groupby(product_descriptor)['target'].sum().reset_index(name = 'sum_target')
product_target_count = train.groupby(product_descriptor)['target'].count().reset_index(name = 'count_target')
product_target_stats = pd.merge(product_target_sum,product_target_count,on=product_descriptor)

train = train.merge(product_target_stats,on=product_descriptor,how='left')
test = test.merge(product_target_stats,on=product_descriptor,how='left')

train['mean_target'] = (train['sum_target'] - train['target'])/(train['count_target']-1)
test['mean_target'] = test['sum_target']/test['count_target']


    train.drop(['count_target','sum_target'],axis=1) #这块需要指定inplace 没指定相当于打印不改变DF
test.drop(['count_target','sum_target'],axis=1).head()
    IDmonthsku_hashproduct_typeproduct_gendermacro_functionfunctionsub_functionmodelaesthetic_sub_linemacro_materialcolorfr_FR_priceen_US_descriptionmean_target
    0ed4c7471eac7e8c6e6718364c2b6e75462eeb47c_11ed4c7471eac7e8c6e6718364c2b6e75462eeb47cAccessoriesWomenTEXTILESSCARVESWINTER SCARVESECHARPE REYKJAVIKSANS LIGNE ESTHETIQUETEXTILENaN650.0In pure, soft cashmere, the Reykjavik scarf is...782.0
    126b2c4f6281482cccf1e748ef388f1649ecf1c8b_1126b2c4f6281482cccf1e748ef388f1649ecf1c8bLeather GoodsWomenCITY BAGSDAILY BAGSHANDBAGSVERY ZIPPED TOTEVERYLEATHERKAKI FANGO COQUILLE2550.0For Fall-Winter 2018, the Very Zipped Tote han...179.0
    200aad36ebc04aaa761df219c1b48d1e6162d28bb_1100aad36ebc04aaa761df219c1b48d1e6162d28bbLeather GoodsWomenCITY BAGSDAILY BAGSCROSS SHOULDER BAGSHINAMAHINALEATHERNOIR2460.0The Hina is two bags in one: gather in the sid...NaN
    38281636e7d162c92c1a3c1f9e40afb05d39a02f9_118281636e7d162c92c1a3c1f9e40afb05d39a02f9AccessoriesMenTEXTILESSTOLESWINTER STOLESETOLE MONOGRAM NIGHTFALLSANS LIGNE ESTHETIQUETEXTILENaN455.0With its subtle tone-on-tone Monogram pattern ...NaN
    4b76f8505700409a198fe69fb33e5d6cc01166c9b_11b76f8505700409a198fe69fb33e5d6cc01166c9bAccessoriesMenTIESTIESOTHER TIESCRAVATE MONOGRAM ILLUSIONSANS LIGNE ESTHETIQUETEXTILENaN150.0Suitable both for business and special occasio...NaN
    train.head()
    IDmonthsku_hashproduct_typeproduct_gendermacro_functionfunctionsub_functionmodelaesthetic_sub_linemacro_materialcolorfr_FR_priceen_US_descriptiontargetsum_targetcount_targetmean_target
    03d8a4ae769b526187c36901f204691a663333fa4_113d8a4ae769b526187c36901f204691a663333fa4Leather GoodsWomenLG ACCESSORIESSMALL LEATHER GOODSCOMPACT WALLETSPF.VICTORINEDAMIER AZURCANVASAZUR370.0Fashioned from summery Damier Azur canvas, thi...1366.01837.02471.000000
    1c05a54f7067be054ec4b27d0d6081353ef7d9df6_11c05a54f7067be054ec4b27d0d6081353ef7d9df6AccessoriesWomenFANCY ACCESSORIESBAG CHARMSBAG CHARMSBAG CHARM PETITE MALLESANS LIGNE ESTHETIQUECANVASNaN295.0This collectible piece features our iconic Pet...526.0526.01NaN
    27cb4d3626bd48a9b523d8693266219c34aeccde8_117cb4d3626bd48a9b523d8693266219c34aeccde8Leather GoodsWomenCITY BAGSDAILY BAGSSHOULDER BAGSGIROLATAMAHINALEATHERGALET2660.0The perfect summer companion, this effortlessl...503.01395.04297.333333
    38e5967f90ff9fb93aa2840fceecc537b2e8e6ff3_118e5967f90ff9fb93aa2840fceecc537b2e8e6ff3Leather GoodsWomenCITY BAGSDAILY BAGSSHOULDER BAGSGIROLATAMAHINALEATHERMAGNOLIA2660.0The perfect summer companion, this effortlessl...276.01395.04373.000000
    46913a128945e0efeafc52101dcdeaa610eaa4430_116913a128945e0efeafc52101dcdeaa610eaa4430AccessoriesMenFASHION JEWELRYNECKLACESNECKLACESCOLL.CHARMS FOR GENTLEMENSANS LIGNE ESTHETIQUENON PRECIOUS METALNaN325.0This distinctive charm necklace reinterprets d...206.0206.01NaN
    train.drop(['count_target','sum_target'],axis=1,inplace =True) 
test.drop(['count_target','sum_target'],axis=1,inplace =True)

    train.head()
    IDmonthsku_hashproduct_typeproduct_gendermacro_functionfunctionsub_functionmodelaesthetic_sub_linemacro_materialcolorfr_FR_priceen_US_descriptiontargetmean_target
    03d8a4ae769b526187c36901f204691a663333fa4_113d8a4ae769b526187c36901f204691a663333fa4Leather GoodsWomenLG ACCESSORIESSMALL LEATHER GOODSCOMPACT WALLETSPF.VICTORINEDAMIER AZURCANVASAZUR370.0Fashioned from summery Damier Azur canvas, thi...1366.0471.000000
    1c05a54f7067be054ec4b27d0d6081353ef7d9df6_11c05a54f7067be054ec4b27d0d6081353ef7d9df6AccessoriesWomenFANCY ACCESSORIESBAG CHARMSBAG CHARMSBAG CHARM PETITE MALLESANS LIGNE ESTHETIQUECANVASNaN295.0This collectible piece features our iconic Pet...526.0NaN
    27cb4d3626bd48a9b523d8693266219c34aeccde8_117cb4d3626bd48a9b523d8693266219c34aeccde8Leather GoodsWomenCITY BAGSDAILY BAGSSHOULDER BAGSGIROLATAMAHINALEATHERGALET2660.0The perfect summer companion, this effortlessl...503.0297.333333
    38e5967f90ff9fb93aa2840fceecc537b2e8e6ff3_118e5967f90ff9fb93aa2840fceecc537b2e8e6ff3Leather GoodsWomenCITY BAGSDAILY BAGSSHOULDER BAGSGIROLATAMAHINALEATHERMAGNOLIA2660.0The perfect summer companion, this effortlessl...276.0373.000000
    46913a128945e0efeafc52101dcdeaa610eaa4430_116913a128945e0efeafc52101dcdeaa610eaa4430AccessoriesMenFASHION JEWELRYNECKLACESNECKLACESCOLL.CHARMS FOR GENTLEMENSANS LIGNE ESTHETIQUENON PRECIOUS METALNaN325.0This distinctive charm necklace reinterprets d...206.0NaN
  • 您还可以看一下 李月喜老师的企业微信开发自建内部应用开发篇课程中的 开发文档如何阅读,如何寻求帮助,如何找到同行小节, 巩固相关知识点