AIM-PIbd-31-Anisin-R-S/lab_7/lab7.ipynb

Лабораторная 7¶

Датасет: Набор данных для анализа и прогнозирования сердечного приступа

In [2]:

import pandas as pd

df = pd.read_csv(".//static//csv//heart_2020_cleaned.csv")

df

Out[2]:

<style scoped=""> .dataframe tbody tr th:only-of-type { vertical-align: middle; } .dataframe tbody tr th { vertical-align: top; } .dataframe thead th { text-align: right; } </style>

	HeartDisease	BMI	Smoking	AlcoholDrinking	Stroke	PhysicalHealth	MentalHealth	DiffWalking	Sex	AgeCategory	Race	Diabetic	PhysicalActivity	GenHealth	SleepTime	Asthma	KidneyDisease	SkinCancer
0	No	16.60	Yes	No	No	3.0	30.0	No	Female	55-59	White	Yes	Yes	Very good	5.0	Yes	No	Yes
1	No	20.34	No	No	Yes	0.0	0.0	No	Female	80 or older	White	No	Yes	Very good	7.0	No	No	No
2	No	26.58	Yes	No	No	20.0	30.0	No	Male	65-69	White	Yes	Yes	Fair	8.0	Yes	No	No
3	No	24.21	No	No	No	0.0	0.0	No	Female	75-79	White	No	No	Good	6.0	No	No	Yes
4	No	23.71	No	No	No	28.0	0.0	Yes	Female	40-44	White	No	Yes	Very good	8.0	No	No	No
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
319790	Yes	27.41	Yes	No	No	7.0	0.0	Yes	Male	60-64	Hispanic	Yes	No	Fair	6.0	Yes	No	No
319791	No	29.84	Yes	No	No	0.0	0.0	No	Male	35-39	Hispanic	No	Yes	Very good	5.0	Yes	No	No
319792	No	24.24	No	No	No	0.0	0.0	No	Female	45-49	Hispanic	No	Yes	Good	6.0	No	No	No
319793	No	32.81	No	No	No	0.0	0.0	No	Female	25-29	Hispanic	No	No	Good	12.0	No	No	No
319794	No	46.56	No	No	No	0.0	0.0	No	Female	80 or older	Hispanic	No	Yes	Good	8.0	No	No	No

319795 rows × 18 columns

Создадим лингвистические переменные и произведем их настройку

Входные: BMI (индекс массы тела) и SleepTime (продолжительность сна)

Выходные: PhysicalHealth (физическое здоровье)

In [3]:

import numpy as np
import skfuzzy as fuzz
from skfuzzy import control as ctrl

bmi = ctrl.Antecedent(np.arange(df["BMI"].min(), df["BMI"].max() + 0.1, 0.1), "BMI")
sleep_time = ctrl.Antecedent(np.arange(df["SleepTime"].min(), df["SleepTime"].max() + 0.1, 0.1), "SleepTime")
physical_health = ctrl.Consequent(np.arange(df["PhysicalHealth"].min(), df["PhysicalHealth"].max() + 1, 1), "PhysicalHealth")

bmi["low"] = fuzz.zmf(bmi.universe, 12, 18)
bmi["normal"] = fuzz.trapmf(bmi.universe, [16, 19, 24, 27])
bmi["high"] = fuzz.smf(bmi.universe, 25, 35)

sleep_time["short"] = fuzz.zmf(sleep_time.universe, 3, 5)
sleep_time["normal"] = fuzz.trapmf(sleep_time.universe, [4, 6, 8, 9])
sleep_time["long"] = fuzz.smf(sleep_time.universe, 8, 10)

physical_health["good"] = fuzz.zmf(physical_health.universe, 0, 5)
physical_health["moderate"] = fuzz.trapmf(physical_health.universe, [3, 7, 15, 20])
physical_health["poor"] = fuzz.smf(physical_health.universe, 15, 30)

bmi.view()
sleep_time.view()
physical_health.view()

c:\Users\User\Desktop\aim\aimvenv\Lib\site-packages\skfuzzy\control\fuzzyvariable.py:125: UserWarning: FigureCanvasAgg is non-interactive, and thus cannot be shown
  fig.show()

Сформируем базу нечетких правил

In [4]:

rule1 = ctrl.Rule(bmi["low"] & sleep_time["short"], physical_health["moderate"])
rule2 = ctrl.Rule(bmi["low"] & sleep_time["normal"], physical_health["good"])
rule3 = ctrl.Rule(bmi["low"] & sleep_time["long"], physical_health["good"])

rule4 = ctrl.Rule(bmi["normal"] & sleep_time["short"], physical_health["moderate"])
rule5 = ctrl.Rule(bmi["normal"] & sleep_time["normal"], physical_health["good"])
rule6 = ctrl.Rule(bmi["normal"] & sleep_time["long"], physical_health["good"])

rule7 = ctrl.Rule(bmi["high"] & sleep_time["short"], physical_health["poor"])
rule8 = ctrl.Rule(bmi["high"] & sleep_time["normal"], physical_health["moderate"])
rule9 = ctrl.Rule(bmi["high"] & sleep_time["long"], physical_health["moderate"])

rule1.view()

Out[4]:

(<Figure size 640x480 with 1 Axes>, <Axes: >)

Создадим нечеткую систему и добавим созданные нечеткие правила в ее базу знаний

In [5]:

health_ctrl = ctrl.ControlSystem(
    [
        rule1,
        rule2,
        rule3,
        rule4,
        rule5,
        rule6,
        rule7,
        rule8,
        rule9
    ]
)

health_simulation = ctrl.ControlSystemSimulation(health_ctrl)

health_ctrl.view()

c:\Users\User\Desktop\aim\aimvenv\Lib\site-packages\skfuzzy\control\controlsystem.py:135: UserWarning: FigureCanvasAgg is non-interactive, and thus cannot be shown
  fig.show()

Проверка расчета выходной переменной

In [6]:

health_simulation.input["BMI"] = 26
health_simulation.input["SleepTime"] = 7

health_simulation.compute()

health_simulation.print_state()

health_simulation.output["PhysicalHealth"]

=============
 Antecedents 
=============
Antecedent: BMI                     = 26
  - low                             : 0.0
  - normal                          : 0.3333333333333333
  - high                            : 0.02003199999999994
Antecedent: SleepTime               = 7
  - short                           : 0.0
  - normal                          : 1.0
  - long                            : 0.0

=======
 Rules 
=======
RULE #0:
  IF BMI[low] AND SleepTime[short] THEN PhysicalHealth[moderate]
	AND aggregation function : fmin
	OR aggregation function  : fmax

  Aggregation (IF-clause):
  - BMI[low]                                               : 0.0
  - SleepTime[short]                                       : 0.0
                             BMI[low] AND SleepTime[short] = 0.0
  Activation (THEN-clause):
                                  PhysicalHealth[moderate] : 0.0

RULE #1:
  IF BMI[low] AND SleepTime[normal] THEN PhysicalHealth[good]
	AND aggregation function : fmin
	OR aggregation function  : fmax

  Aggregation (IF-clause):
  - BMI[low]                                               : 0.0
  - SleepTime[normal]                                      : 1.0
                            BMI[low] AND SleepTime[normal] = 0.0
  Activation (THEN-clause):
                                      PhysicalHealth[good] : 0.0

RULE #2:
  IF BMI[low] AND SleepTime[long] THEN PhysicalHealth[good]
	AND aggregation function : fmin
	OR aggregation function  : fmax

  Aggregation (IF-clause):
  - BMI[low]                                               : 0.0
  - SleepTime[long]                                        : 0.0
                              BMI[low] AND SleepTime[long] = 0.0
  Activation (THEN-clause):
                                      PhysicalHealth[good] : 0.0

RULE #3:
  IF BMI[normal] AND SleepTime[short] THEN PhysicalHealth[moderate]
	AND aggregation function : fmin
	OR aggregation function  : fmax

  Aggregation (IF-clause):
  - BMI[normal]                                            : 0.3333333333333333
  - SleepTime[short]                                       : 0.0
                          BMI[normal] AND SleepTime[short] = 0.0
  Activation (THEN-clause):
                                  PhysicalHealth[moderate] : 0.0

RULE #4:
  IF BMI[normal] AND SleepTime[normal] THEN PhysicalHealth[good]
	AND aggregation function : fmin
	OR aggregation function  : fmax

  Aggregation (IF-clause):
  - BMI[normal]                                            : 0.3333333333333333
  - SleepTime[normal]                                      : 1.0
                         BMI[normal] AND SleepTime[normal] = 0.3333333333333333
  Activation (THEN-clause):
                                      PhysicalHealth[good] : 0.3333333333333333

RULE #5:
  IF BMI[normal] AND SleepTime[long] THEN PhysicalHealth[good]
	AND aggregation function : fmin
	OR aggregation function  : fmax

  Aggregation (IF-clause):
  - BMI[normal]                                            : 0.3333333333333333
  - SleepTime[long]                                        : 0.0
                           BMI[normal] AND SleepTime[long] = 0.0
  Activation (THEN-clause):
                                      PhysicalHealth[good] : 0.0

RULE #6:
  IF BMI[high] AND SleepTime[short] THEN PhysicalHealth[poor]
	AND aggregation function : fmin
	OR aggregation function  : fmax

  Aggregation (IF-clause):
  - BMI[high]                                              : 0.02003199999999994
  - SleepTime[short]                                       : 0.0
                            BMI[high] AND SleepTime[short] = 0.0
  Activation (THEN-clause):
                                      PhysicalHealth[poor] : 0.0

RULE #7:
  IF BMI[high] AND SleepTime[normal] THEN PhysicalHealth[moderate]
	AND aggregation function : fmin
	OR aggregation function  : fmax

  Aggregation (IF-clause):
  - BMI[high]                                              : 0.02003199999999994
  - SleepTime[normal]                                      : 1.0
                           BMI[high] AND SleepTime[normal] = 0.02003199999999994
  Activation (THEN-clause):
                                  PhysicalHealth[moderate] : 0.02003199999999994

RULE #8:
  IF BMI[high] AND SleepTime[long] THEN PhysicalHealth[moderate]
	AND aggregation function : fmin
	OR aggregation function  : fmax

  Aggregation (IF-clause):
  - BMI[high]                                              : 0.02003199999999994
  - SleepTime[long]                                        : 0.0
                             BMI[high] AND SleepTime[long] = 0.0
  Activation (THEN-clause):
                                  PhysicalHealth[moderate] : 0.0


==============================
 Intermediaries and Conquests 
==============================
Consequent: PhysicalHealth           = 3.9601889413098417
  good:
    Accumulate using accumulation_max : 0.3333333333333333
  moderate:
    Accumulate using accumulation_max : 0.02003199999999994
  poor:
    Accumulate using accumulation_max : 0.0

Out[6]:

np.float64(3.9601889413098417)

In [7]:

physical_health.view(sim=health_simulation)

Сравним результаты работы системы с реальными данными

In [8]:

def fuzzy_pred(row):
    health_simulation.input["BMI"] = row["BMI"]
    health_simulation.input["SleepTime"] = row["SleepTime"]
    health_simulation.compute()
    return health_simulation.output["PhysicalHealth"]

result = df.copy()

result["PhysicalHealth_pred"] = result.apply(fuzzy_pred, axis=1)

result.loc[:20, ["BMI", "SleepTime", "PhysicalHealth", "PhysicalHealth_pred"]]

Out[8]:

<style scoped=""> .dataframe tbody tr th:only-of-type { vertical-align: middle; } .dataframe tbody tr th { vertical-align: top; } .dataframe thead th { text-align: right; } </style>

	BMI	SleepTime	PhysicalHealth	PhysicalHealth_pred
0	16.60	5.0	3.0	2.047325
1	20.34	7.0	0.0	1.490667
2	26.58	8.0	20.0	7.808681
3	24.21	6.0	0.0	1.505537
4	23.71	8.0	28.0	1.490667
5	28.87	12.0	6.0	11.427253
6	21.63	4.0	15.0	11.381356
7	31.64	9.0	5.0	11.381356
8	26.45	5.0	0.0	6.758991
9	40.69	10.0	0.0	11.280000
10	34.30	15.0	30.0	11.281767
11	28.71	5.0	0.0	11.432978
12	28.37	8.0	0.0	11.444414
13	28.15	7.0	7.0	11.451294
14	29.29	5.0	0.0	11.411281
15	29.18	6.0	1.0	11.415595
16	26.26	10.0	5.0	5.424223
17	22.59	8.0	0.0	1.490667
18	29.86	5.0	0.0	11.387517
19	18.13	8.0	0.0	1.616988
20	21.16	8.0	0.0	1.490667

Оценим результаты с помощью метрик регрессии

In [9]:

import math
from sklearn import metrics

rmetrics = {}
rmetrics["RMSE"] = math.sqrt(
    metrics.mean_squared_error(result["PhysicalHealth"], result["PhysicalHealth_pred"])
)
rmetrics["RMAE"] = math.sqrt(
    metrics.mean_absolute_error(result["PhysicalHealth"], result["PhysicalHealth_pred"])
)
rmetrics["R2"] = metrics.r2_score(
    result["PhysicalHealth"], result["PhysicalHealth_pred"]
)

rmetrics

Out[9]:

{'RMSE': 9.902693515392837,
 'RMAE': 2.8027515579331546,
 'R2': -0.5512467826513943}

Как можно заметить, нечеткая система с такими переменными и настройками со своей задачей справляется плохо