Fit and Predict

Part of the Supervised Learning section of Coddy's Machine Learning journey. Lesson 3 of 58.

Every scikit-learn model works in two calls. fit studies the training rows with their labels; predict labels rows the model has never seen:

A diagram: six pets, X_train with weight and ear length beside y_train, three cats and three dogs, go through fit into a model. Two new pets, X_new, go through predict, using the model, and come out as cat and dog

The model from this journey's first lesson is k-nearest neighbors: it labels a row like the training rows closest to it. Pets by weight (kg) and ear length (cm):

from sklearn.neighbors import KNeighborsClassifier

X_train = np.array([[4, 6], [5, 7], [3, 5], [30, 10], [25, 12], [35, 11]])
y_train = np.array(["cat", "cat", "cat", "dog", "dog", "dog"])
model = KNeighborsClassifier()
model.fit(X_train, y_train)
print(model.predict([[6, 6], [28, 11]]))

This prints ['cat' 'dog']. New rows go in as a table, one row per pet.

Make, fit, predict: every model in this section follows these three steps.

Challenge

Easy

A birdwatching club's app names birds from wingspan and weight. The starter splits the club's birds.csv and reads tonight's birds into the table spotted.

Make a k-nearest neighbors model and fit it on the training birds.
Predict every spotted bird in one call.
Print one line per bird, as below.

For three birds on another walk:

Bird 1: sparrow
Bird 2: blackbird
Bird 3: robin

Try it yourself

import pandas as pd
from sklearn.model_selection import train_test_split

# One row per bird the club has named: wingspan in cm, weight in grams
birds = pd.read_csv("birds.csv")
X = birds[["wingspan_cm", "weight_g"]]
y = birds["species"]
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.25, random_state=0)

# Tonight's walk: how many birds, then each one's wingspan and weight
count = int(input())
spans = []
weights = []
for i in range(count):
    spans.append(float(input()))
    weights.append(float(input()))
spotted = pd.DataFrame({"wingspan_cm": spans, "weight_g": weights})

# Write your code below