I have an set of items, each with an age (measured in days). For each age I wish to calculate a level of importance (weight) that will be used in a following function.
The younger the age the larger the weight (more recent items are more influential and carry more weight).
I don't wish for the age/weight relationship to be linear. The weight should drop off more steeply as the age increases.
What formula would generate a weight value when passed in an age value?
How can I control the 'drop-off' rate of weight as the age increases (the curve of the graph)?
Simple language and explanations rather than lots of formula would be appreciated by this novice :-)
UPDATE: Using Tanner's assumptions to more specifically ask the question:
The weight drops off more steeply as the age increases, the weight will eventually hit zero. The age value that will produce a weight of zero will be called A.
When the age is zero I know what the weight will be; I call this weight W.
d is the initial drop-off rate: at the beginning, if the age increases by a small value (say 1), the weight will decrease proportionally by d.
I'm not allowed to post images yet but you can view a graph mockup of what I'm attempting at https://img.skitch.com/20120307-bxu8c6t2crubq6c59kprwgcdrc.png