Predicting Numbers

Part of the Supervised Learning section of Coddy's Machine Learning journey. Lesson 11 of 58.

This chapter's models predict numbers. By its end, a chart of a model's misses will expose what it overlooked:

A chart of a model's misses against the temperature: the dots arch from -280 up to +140 and back down, around a dashed line at 0, labelled: the model overlooked a curve

Start with a line you know. LinearRegression learns it from a table X, one row per sample and one column per feature:

from sklearn.linear_model import LinearRegression

area = np.array([60, 75, 90, 110, 130, 150])  # houses, in m²
price = np.array([215, 248, 290, 335, 385, 428])  # thousand dollars
X = area.reshape(-1, 1)  # 6 rows, 1 column
model = LinearRegression()
model.fit(X, price)
print(np.round(model.coef_, 2), np.round(model.intercept_, 2))
print(np.round(np.polyfit(area, price, 1), 2))

This prints the same line twice:

[2.39] 71.59
[ 2.39 71.59]

Six houses as dots, price against area, with the thick orange LinearRegression line and the dashed black np.polyfit line lying exactly on top of each other: price = 2.39 * area + 71.59

coef_ holds one slope per feature: each extra m² adds 2.39 thousand dollars. intercept_ is the line's start, at 0 m². Then model.predict([[100]]) prices 100 m² at about 310.85 thousand.

Challenge

Medium

A print shop never published its prices, but you have 10 past jobs: pages and price in dollars. The program reads the pages of a new job.

Fit a LinearRegression on the pages, as a one-column table.
Print the price per page (3 decimals) and the setup fee, the intercept (2 decimals).
Print the new job's predicted price (2 decimals).

For another shop and 120:

Per page: 0.095 dollars
Setup fee: 4.1 dollars
120 pages: 15.5 dollars

Try it yourself

import numpy as np
from sklearn.linear_model import LinearRegression

# 10 past print jobs: pages printed and the price paid, in dollars
pages = np.array([40, 75, 120, 160, 210, 260, 300, 380, 450, 520])
price = np.array([7.38, 12.6, 17.65, 20.69, 26.63, 32.18, 37.76, 46.4, 55.17, 60.89])
new_pages = int(input())

# Write your code below