ML's shap: Based on FIFA 2018 Statistics (2018 Russia World Cup) team match star classification prediction data set using RF random forest + calculating SHAP value single-sample force map/dependency c

ML之shap：基于FIFA 2018 Statistics(20182008 World Cup in Russia)Team match star classification prediction dataset utilizationRF随机森林+计算SHAPValue One-Sample Force Plot/A detailed guide to interpretability by visualizing dependency contribution graphs

目录

基于FIFA 2018 Statistics(20182008 World Cup in Russia)Team match star classification prediction dataset utilizationRF随机森林+计算SHAPValue One-Sample strives to visualize interpretability

 # 1、定义数据集

# 2、数据预处理

# 2.1、分离特征与标签

# 3、Model building and training

# 3.1、数据集切分# 3.2、模型训练

# 4、模型特征重要性解释可视化

# 4.1、A single sample is based onshapvalue for interpretation visualization

# (1)、Pick a piece of sample data and convert toarray格式

# (2)、利用Shap值解释RFC模型

    # T1、基于树模型TreeExplainer创建Explainer并计算SHAP值,And a single-sample effort is visualized(Analyze the interpretation of single-sample predictions)

    # T2、kernel based modelKernelExplainer创建Explainer并计算SHAP值,And a single-sample effort is visualized(Analyze the interpretation of single-sample predictions)

# 4.2、Multiple samples are based onshapvalue for interpretation visualization

# (1)、基于树模型TreeExplainer创建Explainer并计算SHAP值

# (2)、The features of the full validation dataset samplesshap值summary_plot可视化

# (3)、Dependency contribution graphdependence_plot可视化

---

相关文章
ML：Machine Learning InterpretabilitySHAPUnderstanding of value for single-sample single-feature prediction
ML之shap：基于FIFA 2018 Statistics(20182008 World Cup in Russia)Team match star classification prediction dataset utilizationRF随机森林+计算SHAPA detailed walkthrough of the single-sample effort to visualize interpretability
ML之shap：基于FIFA 2018 Statistics(20182008 World Cup in Russia)Team match star classification prediction dataset utilizationRF随机森林+计算SHAPA detailed walkthrough of the single-sample effort to visualize interpretability

基于FIFA 2018 Statistics(20182008 World Cup in Russia)Team match star classification prediction dataset utilizationRF随机森林+计算SHAPValue One-Sample strives to visualize interpretability

 # 1、定义数据集

数据集来源：Dataset：FIFA 2018 Statistics数据集(Predict FIFA 2018 Man of the Match预测2018FIFA's best player of the year)的简介、下载、使用方法之详细攻略_一个处女座的程序猿的博客-CSDN博客

Date	Team	Opponent	Goal Scored	Ball Possession %	Attempts	On-Target	Off-Target	Blocked	Corners	Offsides	Free Kicks	Saves	Pass Accuracy %	Passes	Distance Covered (Kms)	Fouls Committed	Yellow Card	Yellow & Red	Red	Man of the Match	1st Goal	Round	PSO	Goals in PSO	Own goals	Own goal Time
14-06-2018	Russia	Saudi Arabia	5	40	13	7	3	3	6	3	11	0	78	306	118	22	0	0	0	Yes	12	Group Stage	No	0
14-06-2018	Saudi Arabia	Russia	0	60	6	0	3	3	2	1	25	2	86	511	105	10	0	0	0	No		Group Stage	No	0
15-06-2018	Egypt	Uruguay	0	43	8	3	3	2	0	1	7	3	78	395	112	12	2	0	0	No		Group Stage	No	0
15-06-2018	Uruguay	Egypt	1	57	14	4	6	4	5	1	13	3	86	589	111	6	0	0	0	Yes	89	Group Stage	No	0
15-06-2018	Morocco	Iran	0	64	13	3	6	4	5	0	14	2	86	433	101	22	1	0	0	No		Group Stage	No	0	1	90

# 2、数据预处理

# 2.1、分离特征与标签

df_X    Goal Scored  Ball Possession %  Attempts  ...  Yellow & Red  Red  Goals in PSO
0            5                 40        13  ...             0    0             0
1            0                 60         6  ...             0    0             0
2            0                 43         8  ...             0    0             0
3            1                 57        14  ...             0    0             0
4            0                 64        13  ...             0    0             0

[5 rows x 18 columns]
df_y 0     True
1    False
2    False
3     True
4    False
Name: Man of the Match, dtype: bool

# 3、Model building and training

# 3.1、数据集切分
# 3.2、模型训练

# 4、模型特征重要性解释可视化

# 4.1、A single sample is based onshapvalue for interpretation visualization

# (1)、Pick a piece of sample data and convert toarray格式

输出当前测试样本：5 
 Goal Scored                 2
Ball Possession %          38
Attempts                   13
On-Target                   7
Off-Target                  4
Blocked                     2
Corners                     6
Offsides                    1
Free Kicks                 18
Saves                       1
Pass Accuracy %            69
Passes                    399
Distance Covered (Kms)    148
Fouls Committed            25
Yellow Card                 1
Yellow & Red                0
Red                         0
Goals in PSO                3
Name: 118, dtype: int64
输出当前测试样本的真实label： False
输出当前测试样本的的预测概率： [[0.29 0.71]]

输出当前测试样本：7 
 Goal Scored                 0
Ball Possession %          53
Attempts                   16
On-Target                   4
Off-Target                 10
Blocked                     2
Corners                     7
Offsides                    1
Free Kicks                 20
Saves                       1
Pass Accuracy %            77
Passes                    466
Distance Covered (Kms)    107
Fouls Committed            23
Yellow Card                 1
Yellow & Red                0
Red                         0
Goals in PSO                0
Name: 35, dtype: int64
输出当前测试样本的真实label： False
输出当前测试样本的的预测概率： [[0.56 0.44]]

# (2)、利用Shap值解释RFC模型

    # T1、基于树模型TreeExplainer创建Explainer并计算SHAP值,And a single-sample effort is visualized(Analyze the interpretation of single-sample predictions)

 

    # T2、kernel based modelKernelExplainer创建Explainer并计算SHAP值,And a single-sample effort is visualized(Analyze the interpretation of single-sample predictions)

 

# 4.2、Multiple samples are based onshapvalue for interpretation visualization

# (1)、基于树模型TreeExplainer创建Explainer并计算SHAP值

# (2)、The features of the full validation dataset samplesshap值summary_plot可视化

# (3)、Dependency contribution graphdependence_plot可视化