# Thread: Underlying processes of pattern recognition?

1. I was wondering what the processes are that the mind undergoes when attempting to find a pattern? According to the feature analysis model on the wiki page http://en.wikipedia.org/wiki/Pattern_recognition_(psychology) the following occurs:

Detection
Pattern dissection
Feature comparison in memory
Recognition

I would think that "Feature comparison in memory" would be using working memory and reasoning ability (very vague, what processes does reasoning use?)

Might anyone get into these a little more deeply?

2.

3. I've put some more thought into this question and what I think the real core of what I'd like to discover is how is it that we make comparisons between stored stimuli while pattern finding?

4. imagine walking through a forest.
As you walk you look at each tree and determine its properties.
Things like texture, height, width, color, leaf shape...
You plot each one on a chart with a dimension for each property.
After a while you will notice that the points on the plot tend to fall into groups in a way that cant be explained by random chance.
For example you would notice a big difference between conifers and deciduous trees.
Furthermore, each group tends to breakdown into subgroups.
Give each of these groups and subgroups a different name and you can now identify each tree.

Later you will somehow discover that each new subgroup possesses new properties (dimensions) that the higher groups dont have.
You will then begin the process again with the new dimensions leading to still more subgroups.

5. Thanks for the response, granpa! Let me try to be more specific with my question.

Sorry about how vague my question is. Let me attempt to clear it up a bit using an example.

Say for instance we have the following string of numbers:
2, 8, 18, 32, 50, ?

One compares 8 and 2, there's more than likely different ways of doing this, one way is the difference between the two.
8 - 2 = 6
We store the value 6 in our working memory.
Then we go on repeating the same step except with the following:
18 - 8 = 10

One may then compare the differences between the differences:
10-6 = 4
So we see that 32 was a resultant of 18+14, 50 was 32+18 and the next number "?" will logically be 72.

What are the processes which tell us what comparisons to make between the integers? Subtraction certainly won't work for every situation, so what goes on in our mind that we use to differentiate one set of data from another to find a pattern?

6. Well thats harder.
The first thing I would look at is the overall shape of the overall graph
Is it linear?
Is it logarithmic?
Does it fluctuate?

Bear in mind that any continuous function can be described by a taylor series
http://en.wikipedia.org/wiki/Taylor_series

Beyond that I dont really know.

7. Originally Posted by granpa
Well thats harder.
The first thing I would look at is the overall shape of the overall graph
Is it linear?
Is it logarithmic?
Does it fluctuate?

Bear in mind that any continuous function can be described by a taylor series
http://en.wikipedia.org/wiki/Taylor_series

Beyond that I dont really know.
The pattern itself is more difficult to discern or the root of the question that I'm getting at is more difficult to answer?

8. both

9. Originally Posted by granpa
both
Might you be able to direct me to anything I could read or what have you to answer my question?

10. no.
not really.

11. Hi,
I see that you are interested in Pattern recognition so I will give you link where you can find several books about it.
The present books are intended to collect representative researches around the globe focusing on low-level vision, filter design, features and image descriptors, data mining and analysis, and biologically inspired algorithms.

12. It's very importaint to include:

- compulsive preception.

- psycosis/skitzoprenia.

- group think/flock instinct.

- optical illusions.

13. Originally Posted by Mark1234
Hi,
I see that you are interested in Pattern recognition so I will give you link where you can find several books about it.
The present books are intended to collect representative researches around the globe focusing on low-level vision, filter design, features and image descriptors, data mining and analysis, and biologically inspired algorithms.