app.py

# -*- coding: utf-8 -*-
"""python sklearn.ipynb

Automatically generated by Colab.

Original file is located at
    https://colab.research.google.com/drive/1VQQuE7zemTSAUntTrIrPMGhThj4cBH8N
"""

import sklearn

"""# sklearn documentation database"""

from sklearn import datasets

diabetes = datasets.load_diabetes()
diabetes

import pandas as pd

diabetes = pd.read_csv('https://gist.githubusercontent.com/davidneves11/944edb5ecb7bf6d1770eae91cb20d049/raw/50d3d054185815b0c49561f94badedc06ef3c313/diabetes.csv')
diabetes.head()

x = diabetes['idade']
y = diabetes['resultado']

from sklearn.model_selection import train_test_split

x_treino, x_teste, y_treino, y_teste = train_test_split(x, y) # modelo está como default por enquanto, então fica 25% para treino e 75% para teste

x = diabetes.drop('resultado', axis=1)
x

y = diabetes['resultado']
y

from sklearn.model_selection import train_test_split

SEED = 4121988

x_treino, x_teste, y_treino, y_teste = train_test_split(x, y)

from sklearn.tree import DecisionTreeClassifier

clf_arvore = DecisionTreeClassifier(random_state=SEED, max_depth=3)

clf_arvore.fit(x_treino, y_treino)

clf_arvore.score(x_teste, y_teste)

from sklearn.dummy import DummyClassifier

clf_dummy = DummyClassifier(strategy='most_frequent')
clf_dummy.fit(x_treino, y_treino)

clf_dummy.score(x_teste, y_teste)

from sklearn import tree
import matplotlib.pyplot as plt

fig, ax = plt.subplots(figsize=(15, 10))
tree.plot_tree(clf_arvore,
               ax=ax,
               fontsize=10,
               rounded=True,
               filled=True,
               feature_names=x_treino.columns,
               class_names=['Não Ativado', 'Ativado'],)

plt.show

import pandas as pd

batimentos = pd.read_csv('https://gist.githubusercontent.com/davidneves11/d72e7f49ab01c856acc5d07be4b1a9dd/raw/37631e3a40da92e6261c00fffdf0fb9b869b35dd/batimentos%2520cardiacos.csv')
batimentos.head()

import matplotlib.pyplot as plt
import seaborn as sns

sns.set()
plt.figure(figsize=(100, 50))

sns.lmplot(x='Peso', y='Batimentos cardiacos', data=batimentos,line_kws={'color':'red'} )
plt.show()

x = batimentos[['Peso', 'Idade']]

y = batimentos['Batimentos cardiacos']

from sklearn.model_selection import train_test_split

SEED = 4500

x_treino, x_teste, y_treino, y_teste = train_test_split(x, y, random_state=SEED)

from sklearn import linear_model



scaler = StandardScaler()
x_treino_scaled = scaler.fit_transform(x_treino)
x_teste_scaled = scaler.transform(x_teste)


rgs = linear_model.LinearRegression(fit_intercept=False)


rgs.fit(x_treino_scaled, y_treino)


y_pred = rgs.predict(x_teste_scaled)


print("Coeficientes do modelo:", rgs.coef_)

import pandas as pd

colesterol = pd.read_csv('https://gist.githubusercontent.com/davidneves11/01b2963f7a8dfd87d79010fbf847b221/raw/685870f4365bcda4e5bb9e342285e0aac37dd556/colesterol.csv')
colesterol.head()

import seaborn as sns

sns.scatterplot(x = 'pressao_sanguinea_repouso', y = 'colesterol', data = colesterol)

from sklearn.cluster import KMeans

kmeans = KMeans(n_clusters=2, random_state=9)

x = colesterol.select_dtypes('int64')

colesterol.info()

kmeans.fit(x)

kmeans.labels_   # separa em dois grupos (1 e 0)

sns.scatterplot(x = 'pressao_sanguinea_repouso', y = 'colesterol', data = colesterol, hue = kmeans.labels_) # visualização desses "dois" grupos, 1 e 0

colesterol['clusters'] = kmeans.labels_   # identifica a média desses clusters de colesterol
colesterol.head()

colesterol.groupby('clusters')['colesterol'].mean()  # Agrupamento da coluna de clusters baseado na média de colesterol