Mixed Columns

Part of the Supervised Learning section of Coddy's Machine Learning journey. Lesson 46 of 58.

Most tables mix words and numbers, and each kind needs its own step. A ColumnTransformer sends each column to its step and joins the results:

A diagram titled 4 columns in, 7 columns out. Three vinyl records with genre, condition, price and year. Genre and condition go through OneHotEncoder into five 0/1 columns: jazz, rock, good, mint, worn. Price and year go through StandardScaler into two scaled columns

For a record shop, in code:

from sklearn.compose import ColumnTransformer
from sklearn.preprocessing import OneHotEncoder

prep = ColumnTransformer([
    ("words", OneHotEncoder(), ["genre", "condition"]),
    ("numbers", StandardScaler(), ["price", "year"]),
])
pipe = make_pipeline(prep, LogisticRegression(max_iter=1000))

Each entry is a name, a step and its columns. Unlike get_dummies, the encoder remembers the training categories, so one new record gets the same 7 columns.

Challenge

Medium

A chess club wants each player's win chance before a game. The starter splits chess.csv and reads tonight's game.

Encode opening and color, scale rating_gap and minutes, and print how many columns that makes from the training rows.
Chain it with logistic regression, fit, and print the test accuracy.
Print tonight's chance of a win.

Another club, three openings, and sicilian, black, 50, 15:

Columns after preparing: 7
Test accuracy: 0.805
Win chance: 0.66

Try it yourself

import pandas as pd
from sklearn.model_selection import train_test_split
from sklearn.compose import ColumnTransformer
from sklearn.preprocessing import OneHotEncoder, StandardScaler
from sklearn.linear_model import LogisticRegression
from sklearn.pipeline import make_pipeline

# A chess club's games: the opening, the player's color, their rating minus
# the opponent's, minutes on the clock, and whether the player won (1) or not (0)
games = pd.read_csv("chess.csv")
X = games[["opening", "color", "rating_gap", "minutes"]]
y = games["won"]
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.25, random_state=0)

# Tonight's game, one row with the same columns
game = pd.DataFrame({"opening": [input()], "color": [input()],
                     "rating_gap": [int(input())], "minutes": [int(input())]})

# Write your code below