Regression

You can use Hollerith to predict a numerical value such as a price, a duration, or a measurement.

Getting started

The following is a full example using the Hollerith Regressor. This is a regression problem where we aim to predict the compressive strength of concrete from its manufacturing process.

from hollerith import Hollerith, read_csv

df = read_csv("concrete-strength.csv")
train = df[df["__hollerith_split"] == "train"].drop(columns="__hollerith_split")
score = df[df["__hollerith_split"] == "test"].drop(
    columns=["__hollerith_split", "strength_mpa"]
)

reg = Hollerith()
reg.fit(train, target="strength_mpa")

preds = reg.predict(score)
print(preds[:5])
# [13.76, 50.3, 23.37, 32.89, 60.22]

Prediction intervals

Quantile regression returns the range around each prediction. Use it to return low, typical, and high estimates around a predicted value.

result = reg.predict(score, quantiles=[0.1, 0.5, 0.9])
result.head(1)
#    prediction    0.1    0.5    0.9
# 0       13.76  10.13  14.58  15.28

Each quantile column estimates different points on the model’s predicted distribution of the target. You can request a denser approximation if useful: [0.01, 0.05, 0.1, 0.25, 0.5, 0.75, 0.9, 0.95, 0.99].

Evaluating accuracy

Learn how Hollerith scores itself

Classification

Learn about Hollerith’s classification capabilities

Get started

Guides

Reference

Account

Help

Getting started

Prediction intervals

Evaluating accuracy

Classification

​Getting started

​Prediction intervals

​Related

Evaluating accuracy

Classification

Getting started

Prediction intervals

Related