心脏衰竭致死率预测包含随机森林回归、决策树、svm等算法

D:\Program\CONDA\python.exe "D:\software\pycharm\PyCharm\PyCharm 2020.2\plugins\python\helpers\pydev\pydevd.py" --multiproc --qt-support=auto --client 127.0.0.1 --port 49324 --file D:/Program/JacksonProject/machine_learning/heart_beat_death/heart_failure_clinical.py pydev debugger: process 24716 is connecting Connected to pydev debugger (build 202.6397.98) There is 299 observation and 13 columns in the dataset age has 47 unique value anaemia has 2 unique value creatinine_phosphokinase has 208 unique value diabetes has 2 unique value ejection_fraction has 17 unique value high_blood_pressure has 2 unique value platelets has 176 unique value serum_creatinine has 40 unique value serum_sodium has 27 unique value sex has 2 unique value smoking has 2 unique value time has 148 unique value DEATH_EVENT has 2 unique value column name : DEATH_EVENT -------------------------------- per_of_nulls : % 0.0 num_of_nulls : 0 num_of_uniques : 2 0 203 1 96 Name: DEATH_EVENT, dtype: int64 0 203 1 96 Name: DEATH_EVENT, dtype: int64 Percentage of DEATH: % 32.11 --> (96 cases for Heart Disease) Percentage of NOT DEATH: % 67.89 --> (203 cases for NOT Heart Disease) <pandas.io.formats.style.Styler object at 0x000002342FC589E8> Skewness: 0.7703488154389491 Kurtosis: -1.4160799150023284 Numerical Columns: Index(['age', 'anaemia', 'creatinine_phosphokinase', 'diabetes', 'ejection_fraction', 'high_blood_pressure', 'platelets', 'serum_creatinine', 'serum_sodium', 'sex', 'smoking', 'time'], dtype='object') Categorical Columns: Index([], dtype='object') A skewness value of 4.463110084653752 means that the distribution is approx. highly skewed A skewness value of 4.455995882049026 means that the distribution is approx. highly skewed A skewness value of 1.4623208382757793 means that the distribution is approx. highly skewed A skewness value of 0.7703488154389491 means that the distribution is approx. moderately skewed A skewness value of 0.770348815438949 means that the distribution is approx. moderately skewed A skewness value of 0.6267318547287857 means that the distribution is approx. moderately skewed A skewness value of 0.5553827516973211 means that the distribution is approx. moderately skewed A skewness value of 0.42306190672863536 means that the distribution is approx. symmetric A skewness value of 0.3339286842537603 means that the distribution is approx. symmetric A skewness value of 0.2782606644055605 means that the distribution is approx. symmetric A skewness value of 0.12780264559841184 means that the distribution is approx. symmetric A skewness value of -0.6267318547287857 means that the distribution is approx. highly skewed A skewness value of -1.0481360160574988 means that the distribution is approx. highly skewed serum_creatinine 25.828 creatinine_phosphokinase 25.149 dtype: float64 For age and age, there is NO multicollinearity problem For age and anaemia, there is NO multicollinearity problem For age and creatinine_phosphokinase, there is NO multicollinearity problem For age and diabetes, there is NO multicollinearity problem For age and ejection_fraction, there is NO multicollinearity problem For age and high_blood_pressure, there is NO multicollinearity problem For age and platelets, there is NO multicollinearity problem For age and serum_creatinine, there is NO multicollinearity problem For age and serum_sodium, there is NO multicollinearity problem For age and sex, there is NO multicollinearity problem For age and smoking, there is NO multicollinearity problem For age and time, there is NO multicollinearity problem For age and DEATH_EVENT, there is NO multicollinearity problem For anaemia and age, there is NO multicollinearity problem For anaemia and anaemia, there is NO multicollinearity problem For anaemia and creatinine_phosphokinase, there is NO multicollinearity problem For anaemia and diabetes, there is NO multicollinearity problem For anaemia and ejection_fraction, there is NO multicollinearity problem For anaemia and high_blood_pressure, there is NO multicollinearity problem For anaemia and platelets, there is NO multicollinearity problem For anaemia and serum_creatinine, there is NO multicollinearity problem For anaemia and serum_sodium, there is NO multicollinearity problem For anaemia and sex, there is NO multicollinearity problem For anaemia and smoking, there is NO multicollinearity problem For anaemia and time, there is NO multicollinearity problem For anaemia and DEATH_EVENT, there is NO multicollinearity problem For creatinine_phosphokinase and age, there is NO multicollinearity problem For creatinine_phosphokinase and anaemia, there is NO multicollinearity problem For creatinine_phosphokinase and creatinine_phosphokinase, there is NO multicollinearity problem For creatinine_phosphokinase and diabetes, there is NO multicollinearity problem For creatinine_phosphokinase and ejection_fraction, there is NO multicollinearity problem For creatinine_phosphokinase and high_blood_pressure, there is NO multicollinearity problem For creatinine_phosphokinase and platelets, there is NO multicollinearity problem For creatinine_phosphokinase and serum_creatinine, there is NO multicollinearity problem For creatinine_phosphokinase and serum_sodium, there is NO multicollinearity problem For creatinine_phosphokinase and sex, there is NO multicollinearity problem For creatinine_phosphokinase and smoking, there is NO multicollinearity problem For creatinine_phosphokinase and time, there is NO multicollinearity problem For creatinine_phosphokinase and DEATH_EVENT, there is NO multicollinearity problem For diabetes and age, there is NO multicollinearity problem For diabetes and anaemia, there is NO multicollinearity problem For diabetes and creatinine_phosphokinase, there is NO multicollinearity problem For diabetes and diabetes, there is NO multicollinearity problem For diabetes and ejection_fraction, there is NO multicollinearity problem For diabetes and high_blood_pressure, there is NO multicollinearity problem For diabetes and platelets, there is NO multicollinearity problem For diabetes and serum_creatinine, there is NO multicollinearity problem For diabetes and serum_sodium, there is NO multicollinearity problem For diabetes and sex, there is NO multicollinearity problem For diabetes and smoking, there is NO multicollinearity problem For diabetes and time, there is NO multicollinearity problem For diabetes and DEATH_EVENT, there is NO multicollinearity problem For ejection_fraction and age, there is NO multicollinearity problem For ejection_fraction and anaemia, there is NO multicollinearity problem For ejection_fraction and creatinine_phosphokinase, there is NO multicollinearity problem For ejection_fraction and diabetes, there is NO multicollinearity problem For ejection_fraction and ejection_fraction, there is NO multicollinearity problem For ejection_fraction and high_blood_pressure, there is NO multicollinearity problem For ejection_fraction and platelets, there is NO multicollinearity problem For ejection_fraction and serum_creatinine, there is NO multicollinearity problem For ejection_fraction and serum_sodium, there is NO multicollinearity problem For ejection_fraction and sex, there is NO multicollinearity problem For ejection_fraction and smoking, there is NO multicollinearity problem For ejection_fraction and time, there is NO multicollinearity problem For ejection_fraction and DEATH_EVENT, there is NO multicollinearity problem For high_blood_pressure and age, there is NO multicollinearity problem For high_blood_pressure and anaemia, there is NO multicollinearity problem For high_blood_pressure and creatinine_phosphokinase, there is NO multicollinearity problem For high_blood_pressure and diab