})
cp = pd.get_dummies(df["cp"])
df = pd.concat([df, cp], axis=1)

df["slope"] = df["slope"].map({1: "upsloping", 2: "flat", 3: "downsloping"})
slope = pd.get_dummies(df["slope"])
df = pd.concat([df, slope], axis=1)

data = df.loc[:, [
    "normal", "fixed_defect", "reversible defect", "typical angina",
    "atypical angina", "non-anginal pain", "asymptomatic", "upsloping", "flat",
    "downsloping", "exang", "ca", "labels"
]]

features = data.columns.tolist()
features.remove("labels")
heart = ml.ai(data=df, features=features, target="labels", test_size=0.2)

import pickle
pickle.dump(heart.model, open("logregheart.pkl", "wb"))
# dump python file into pickle file
# you determine here
# pkl file appears in the same directory as this python file

# just recall the pickle file back into one name ( which will be model )
# you can test with myheart.coef_
# since they are in the same directory , ypu don't need to direct to it
# just call him by his name
# although you gotta figure out how to call the other directory with . or ..
myheart = pickle.load(open("logregheart.pkl", "rb"))
Exemplo n.º 2
0
thal = pd.get_dummies(df["thal"])
thal.columns = ["normal", "fixed defect", "reversable defect"]
df = pd.concat([df, thal], axis=1)
# cp
df["cp"] = df["cp"].map({
    1: "typical angina",
    2: "atypical angina",
    3: "non-anginal pain",
    4: "asymptomatic"
})
cp = pd.get_dummies(df["cp"])
df = pd.concat([df, cp], axis=1)
# slope
df["slope"] = df["slope"].map({1: "upsloping", 2: "flat", 3: "downsloping"})
slope = pd.get_dummies(df["slope"])
df = pd.concat([df, slope], axis=1)
data = df.loc[:, [
    "normal", "fixed defect", "reversable defect", "typical angina",
    "atypical angina", "non-anginal pain", "asymptomatic", "upsloping", "flat",
    "downsloping", "exang", "ca", "labels"
]]
features = data.columns.tolist()
features.remove("labels")
params = {"dt": {"max_depth": 3}, "lr": {"penalty": "l2"}}
heart = ml.ai(data=df,
              features=features,
              target="labels",
              test_size=0.2,
              model="lr",
              params=params)