AI glossary: terms starting with O

A value in memory, with its type. Every object has an identity number, shown by id().

References to Python objects instead of fixed-width numeric values

Recorded values without assigning the proposed exposure.

B's score minus A's score on the original validation set.

Counts or fractions of recorded values.

p / (1 − p)

Example: 4 for p = 0.8

One pair of different variables in a pair plot.

b, which moves the line without turning it

Example: −4

Choose the simplest setting within one standard error of the best.

A value an operator works on: in 10 + 3, the operands are 10 and 3.

One step we count, here a comparison.

A symbol or word that performs an operation: +, //, ==, and, in.

Categories with an order, such as a kitchen graded excellent to poor.

Chooses the intercept and slope that minimize the summed squared vertical errors.

the point (0, 0)

Example: where the axes cross

perpendicular; dot product zero

Example: (3, 4) and (0.8, −0.6)

trip duration, the recorded quantity the forecast is trying to predict: dropoff time minus pickup time, in minutes.

One combination of assigned action and recorded label, such as a missed default.

An observation far from most others under a stated comparison. Being unusual does not make it wrong.

Fitting the training rows much better than new rows.

a result larger than the largest float, stored as inf

Example: np.exp(710)

The region where two class distributions share values.

Multiple observations cover the same plotted area.

Selecting a declared group at a higher rate by design.