org.opentripplanner.analyst.core

Class WeightingFunction

java.lang.Object

org.opentripplanner.analyst.core.WeightingFunction

Direct Known Subclasses:

WeightingFunction.Logistic, WeightingFunction.SharpCutoff

public abstract class WeightingFunction
extends Object

A weighting function, expressing how influential something is according to its distance from you. Perhaps this could be a FunctionalInterface so implementations can be defined using Java 8 shorthand, but it currently has two functions. The separation into two functions was to allow optimizations for simple hard-cutoff functions, where you can avoid doing a multiplication to determine the final result. I'm not sure eliminating multiplications by 0 and 1 actually has much of an effect on execution speed, so maybe the functional abstraction is more important. But then again, maybe we'll always just use the same few functions, in which case the abstraction is purely to aid comprehension by OTP developers.

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	WeightingFunction.Logistic The logistic function, a commonly used sigmoid (s-shaped function).
static class	WeightingFunction.SharpCutoff The typical hard cutoff used to select all objects within an isochrone curve.

Constructor Summary

Constructors

Constructor and Description
WeightingFunction()

Method Summary

Modifier and Type	Method and Description
abstract int[]	apply(int[] countsPerSecond) Weight the counts (binned by second, so countsPerSecond[i] is destinations reachable in i - i + seconds) and return the output of this weighting function as a _cumulative distribution_.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

WeightingFunction

public WeightingFunction()

Method Detail

apply

public abstract int[] apply(int[] countsPerSecond)

Weight the counts (binned by second, so countsPerSecond[i] is destinations reachable in i - i + seconds) and return the output of this weighting function as a _cumulative distribution_.