Merge pull request #60 from jiangzhonglian/master

更新  设定评估指标 内容

Author: jiangzhonglian

README.md

@@ -141,7 +141,7 @@
 | 深度学习 | 机器学习  | 大数据 | 运维工具 |
 | --- | --- | --- | --- |
 | [TensorFlow R1.2 中文文档](http://cwiki.apachecn.org/pages/viewpage.action?pageId=10030122) | [机器学习实战-教学](https://github.com/apachecn/MachineLearning) | [Spark 2.2.0和2.0.2 中文文档](http://spark.apachecn.org/) | [Zeppelin 0.7.2 中文文档](http://cwiki.apachecn.org/pages/viewpage.action?pageId=10030467) |
-| [Pytorch 3.0 中文文档 ](http://pytorch.apachecn.org/cn/0.3.0/) | [Sklearn 0.19 中文文档](http://sklearn.apachecn.org/) | [Storm 1.1.0和1.0.1 中文文档](http://storm.apachecn.org/) | [Kibana 5.2 中文文档](http://cwiki.apachecn.org/pages/viewpage.action?pageId=8159377) |
+| [Pytorch 0.3 中文文档 ](http://pytorch.apachecn.org/cn/0.3.0/) | [Sklearn 0.19 中文文档](http://sklearn.apachecn.org/) | [Storm 1.1.0和1.0.1 中文文档](http://storm.apachecn.org/) | [Kibana 5.2 中文文档](http://cwiki.apachecn.org/pages/viewpage.action?pageId=8159377) |
 |  | [LightGBM 中文文档](http://lightgbm.apachecn.org/cn/latest) | [Kudu 1.4.0 中文文档](http://cwiki.apachecn.org/pages/viewpage.action?pageId=10813594) |  |
 |  | [XGBoost 中文文档](http://xgboost.apachecn.org/cn/latest)  | [Elasticsearch 5.4 中文文档](http://cwiki.apachecn.org/pages/viewpage.action?pageId=4260364) |
 |  |  | [Beam 中文文档](http://beam.apachecn.org/) |

competitions/featured/mercari-price-suggestion-challenge/README.md

@@ -62,10 +62,60 @@
 
 ## 5.模型评估
 
-* 设定评估指标
+### 设定评估指标
+
+> 1.绝对误差于相对误差
+
+![](/static/images/competitions/featured/mercari-price-suggestion-challenge/EvaluationCriteria/15160975410546.jpg)
+
+> 2.平均绝对误差
+
+![](/static/images/competitions/featured/mercari-price-suggestion-challenge/EvaluationCriteria/15160975616578.jpg)
+
+> 3.均方误差
+
+![](/static/images/competitions/featured/mercari-price-suggestion-challenge/EvaluationCriteria/15160975799651.jpg)
+
+> 4.均方根误差
+
+![](/static/images/competitions/featured/mercari-price-suggestion-challenge/EvaluationCriteria/15160977834397.jpg)
+
+> 5.平均绝对百分误差
+
+![](/static/images/competitions/featured/mercari-price-suggestion-challenge/EvaluationCriteria/15160979383193.jpg)
+
+> 6.Kappa统计
+
+![](/static/images/competitions/featured/mercari-price-suggestion-challenge/EvaluationCriteria/15160982666965.jpg)
+
+> 7.识别准确度（正确率）
+
+![](/static/images/competitions/featured/mercari-price-suggestion-challenge/EvaluationCriteria/15160988495710.jpg)
+
+> 8.识别精确度（ P值 ）
+
+![](/static/images/competitions/featured/mercari-price-suggestion-challenge/EvaluationCriteria/15160989546724.jpg)
+
+> 9.反馈率（ R值 ）
+
+![](/static/images/competitions/featured/mercari-price-suggestion-challenge/EvaluationCriteria/15160989672539.jpg)
+
+> 10.非均衡问题（ F值 ）
+
+
+> 11.ROC曲线/AUC面积
+
+![](/static/images/competitions/featured/mercari-price-suggestion-challenge/EvaluationCriteria/15160989939052.jpg)
+
+> 12.混淆矩阵
+
+![](/static/images/competitions/featured/mercari-price-suggestion-challenge/EvaluationCriteria/15160990207804.jpg)
+
 * 多模型对比
 * 模型优化
 
+
+
 ## 6.模型发布
 
 * 模型部署

src/python/getting-started/digit-recognizer/knn-python3.6.py

@@ -43,7 +43,7 @@ def saveResult(result, csvName):
 
 
 def knnClassify(trainData, trainLabel):
-    knnClf = KNeighborsClassifier(3)   # default:k = 5,defined by yourself:KNeighborsClassifier(n_neighbors=10)
+    knnClf = KNeighborsClassifier()   # default:k = 5,defined by yourself:KNeighborsClassifier(n_neighbors=10)
     knnClf.fit(trainData, ravel(trainLabel)) # ravel Return a contiguous flattened array.
     return knnClf
 

…rted/titanic/titanic-pianke2python3.6.py …ing-started/titanic/titanic-python3.6.pysrc/python/getting-started/titanic/titanic-pianke2python3.6.py renamed to src/python/getting-started/titanic/titanic-python3.6.py src/python/getting-started/titanic/titanic-pianke2python3.6.py renamed to src/python/getting-started/titanic/titanic-python3.6.py

@@ -1,6 +1,6 @@
 # -*- coding: utf-8 -*-
 """
-Created on 2017-12-7 10:38:05 2017
+Created on 2017-12-07
 Update  on 2017-12-24
 Team:   一把梭
 Github: https://github.com/apachecn/kaggle
@@ -17,7 +17,8 @@
 from sklearn.model_selection import cross_val_score
 
 # 0. 数据读入及预处理
-data_train = pd.read_csv('train.csv')
+root_path = 'datasets/getting-started/titanic/input'
+data_train = pd.read_csv('%s/%s' % (root_path, 'train.csv'))
 # data_train.info()
 # print(data_train.describe())