pabloalcain

class Model:
    def __init__(self, number_of_selected=2):
        self.classifier = Pipeline(
            steps=[("select", SelectKBest(k=number_of_selected)),
                   ("clf", LogisticRegression())])
...

pabloalcain

from dataset import Dataset
from model import Model

iris_dataset = Dataset("iris.csv")
model = Model()
model.fit(iris_dataset)
model.save("model.pkl")

pabloalcain

class Model:
    ...
    def fit(self, dataset):
        self.classifier.fit(dataset.features, dataset.target)

pabloalcain

import pandas as pd

class Dataset:
    def __init__(self, filename):
        dataset = pd.read_csv(filename)
        self.features = dataset[["s-l", "s-w", "p-l", "p-w"]]
        self.target = dataset["variety"]

pabloalcain

...
dataset = pd.read_csv("iris.csv")
features = dataset[["s-l", "s-w", "p-l", "p-w"]]
target = dataset["variety"]
...

pabloalcain

class Model:
    def __init__(self):
        self.classifier = Pipeline(
            steps=[("select", SelectKBest(k=2)),
                   ("clf", LogisticRegression())])

    def fit(self, features, target):
        self.classifier.fit(features, target)

    def load(self, filename):

pabloalcain

...
def fit_model(features, target):
    classifier = Pipeline(steps=[("select", SelectKBest(k=2)),
                                 ("clf", LogisticRegression())])
    classifier.fit(features, target)
    return classifier

def load_model(filename):
    ...

pabloalcain

...
classifier = fit_model(features, target)
save_model(classifier, "model.pkl")
...

pabloalcain

#file: train.py
...
classifier = Pipeline(steps=[("select", SelectKBest(k=2)),
                             ("clf", LogisticRegression())])
classifier.fit(features, target)
...

pabloalcain

...
classifier = load_model("model.pkl")
...