Creating an SEO resource using a genetic algorithm

Creating an SEO resource from scratch

[petersmall]

In a thread a few months back, I explained how I've been working on information systems that use strategies found in biological systems.

At the time, it was all very much theoretical stuff and nobody seemed to make much sense of it.

Since then, I've built a practical demo to explain how a biological approach using a genetic algorithm can be used to create a very credible resource for information on search engine optimization.

Now I know this sounds like a lot of old tosh, but if you are familiar with the way genetic algorithms work you may appreciate how they promote good variable values and filter out the bad values.

Applying this principle to links relating to SOE, it is possible to start with a miscellaneous collection of mediocre links and finish up with a very creditable SEO resource that is optimally efficient.

Of course, it's not some magic program that creates a resource at the touch of a button. It requires intelligent human participation and cooperation.

So, if there are any SEO educationalists or practitioners that have a little time to spare and like exploring new ways to process information on SEO, I'd appreciate some help here.

A demo of this process is on my web site.

Peter Small

[IanTurner]

I did some work on genetic algorithms a while back and although I can understand how you think that this approach can be applied, the data set and timescale required to implement would be prohibitively long.

Other Considerations:
The target for the algorithm is constantly moving just to make matters more complex.

Genetic Optimisation algorithms are just as prone to finding false peaks as other mathematical optimisation algorithms.

[petersmall]

Hi Ian,

Thanks for your comments.

In fact it uses the principle of genetic algorithms rather than the rigid mathematical form. It is more like the evolutionary strategy used by natural systems. There isn't a time problem with this.

The nature of the subject matter means there is not an issue with false peaks either, because the aim, like natural systems, is to keep as close to maximum eficiency as a moving target will allow.

Unlike most applications of genetic algorithms, there is no definitive goal to aim for because what is and what is not a good reference for search engine optimization can be subjective, depending upon what type of business it is for. The system allows for the emergence of different kinds of resources to cater for this.

The general idea is that good links are retained by the system and bad links disappear, but because the links are in groups (like items in a genetic algorithm) the bad links do not automatically disappear. This allows for differences of opinion, as to the value of a link.

Anyway, the demo explains most of this.