<DrumBeat: September 7, 2006 The Oil Drum Russian Gas Supplies next winter for Europe ?>

66 comments on The Loglet Analysis

Comments can no longer be added to this story.

| Show without comments | PDF version

Sam Foucher on September 6, 2006 - 6:03pm Permalink | Subthread | Parent | Comments top
Well, the Loglet analysis is a component logistic model, you'll find a good explanation here.  The best way to understand it is to look at the bi-logistic model:

As it turns out, many growth and diffusion processes are actually made up of several subprocesses. First, let us consider the case of a system which experiences growth in two discrete growth phases. Then, we will extend this to an arbitrary number of phases.

Systems with two growth phases follow what we call the ``Bi-logistic'' model [12]. In this model, growth is the sum of two discrete ``wavelets'' , each of which is a three-parameter logistic.


Below is an example of a bi-logistic model:

The cumulative curve in panel A is the sum of the two logistic process in panel B (you then take the derivative to get the usual Hubbert/Bell curves).

The Loglet is an extension of the Bi-logistic case to a multi-logistic case.

Comments can no longer be added to this story.
Boof on September 6, 2006 - 9:22pm Permalink | Subthread | Parent | Comments top
S curves and bell curves

Maybe this requires a bit of a leap of faith. The idea of diminishing returns (more effort, not so much reward) lends itself to S curves. However not every S curve has a bell curve as a tangent.  It has to be just right. A similar example   is using a sheet of shiny metal as a reflector; it has to be an exact parabola to get a sharp focus.

An alternative theory is that the bell curve is really a triangle with fudgy corners. But I don't wanna go there. Given the uncertainties bell curves are OK.

Comments can no longer be added to this story.

66 comments on The Loglet Analysis

Comments can no longer be added to this story.

| Show without comments | PDF version