AI glossary: terms starting with I

Every trial uses the same outcome probabilities.

The name of a variable, function or class. It can use letters, digits and underscores, can not start with a digit, and can not be a keyword. Identifiers are case sensitive.

I, ones on the diagonal and zeros elsewhere; it leaves every vector unchanged

Example: np.eye(2)

Having a large condition number.

Pixel values, one feature per pixel.

An object that can not be changed after it is created: a tuple, a string, a number.

Running a module once and binding a name to it.

Chance that a specified eligible unit enters the sample: 200 / 25,900 in the simple design.

A displayed time bin extending beyond the source's recorded period, such as the week containing 2011-12-09.

Knowing one event does not change the probability of the other.

One trial's outcome does not alter another's probability.

Observations whose joint variation is not tied by the assumed sampling process.

An element's position, counting from zero. A negative index counts from the end, from -1.

the float for infinity

Example: np.exp(1000)

Running a trained model on examples it was not trained on.

A loop that never terminates. In 2.2's example, the condition stays True and there is no other exit. A while True loop with a reachable break can terminate (2.4).

How much a row moves the fit.

Defining a class from another, so that it starts with everything the other has: class VerifiedReview(Review):.

input() shows a prompt and returns what the user types, always as a str.

The values a simulation is allowed to draw under its model: the 334 recorded daily totals of item 85123A.

the value a function takes in, and the one it gives back

Example: 2 in, 5 out

The order keys were first added to a dictionary; updating a value preserves its position.

An object created from a class: review = Review("kettle", 5, "…").

A whole number: 3, -7, 1000.

the limit of the strip sums as the strips get thinner

Example: ∫₀¹ x² dx = 1/3

The product of two features.

Fitted duration at distance zero.

A column of ones whose weight is the intercept.

Executed one statement at a time by a program called the interpreter, so each line's result can be seen straight away.

Every value in both sets, a & b.

Outcomes satisfying both events: 130 rows above 100 and labelled fraud.

A⁻¹, the matrix that undoes A: A⁻¹A = AA⁻¹ = I

Example: [[1, −1], [−1, 2]]

a function that undoes another; its graph is the reflection across y = x

Example: ln undoes exp

Operation that maps a transformed value back to the original scale; here expm1.

Sum of included line quantities times their unit prices for one invoice.

The third quartile minus the first quartile: the spread of the middle half.

Anything a for loop can go through, one item at a time: a list, a string, a range.

One run of the loop body.

a method that improves an answer step by step

Example: bisection, gradient descent

An object that yields its items one at a time when asked for the next item.