Monday, February 18, 2008

I've acquired the data that I need for this project. I have hour-by-hour data for EURUSD for the last nine years. Now I need to find a way to display it in the needed way.

I think I'm going to look for glasstrons.

The Neurex Project

I think I've got a live one here. I'm tired of playing ball. I'm going to use this brain of mine to make money so I don't have to trade my labor for it all the time. I've begun work on something I'm calling "The Neurex Project". The following is some of my thoughts on it so far.

Warning: If you think hypnosis is bullshit, don't bother reading.

Neural Forex

The basis of this project is the ability of a neural network to solve problems.

In this instance, the challenge is to be able to predict with a reasonable degree of accuracy the movement of international currency exchange markets. This happens already, in the form of the "hunches" of seasoned traders. Digital processors also perform these calculations in the form of algorithms, however they are almost always taken with a grain of salt (read: inaccurate).

I suggest that the neural network found between one's ears can be, under the right conditions, trained to perform pattern-analysis on a set of data to the end of predicting future fluctuations. This happens daily, but the subconscious calculations happen as a side-effect of long-term exposure to the data during conscious attempts to do the same work resulting in the aforementioned "hunch".

I believe that it is possible to get a faster and more accurate "hunch generator" by doing away with conscious attempts to interpret the data and exposing the subconscious directly to the data in question - in such a way as to facilitate the analysis and prediction process.

So far I've come up with a few points that need to be addressed:

Raw Data

The neural net will require massive amounts of raw data from which to extract the "black box" algorithm necessary for accuracy. This data must be presented in such a way as to be easily assimilatable and also to facilitate the output of prediction data. This is complicated somewhat by the specific qualities of forex data, such as timescale, trade volume, and currency pair.

Historical data will not only need to be obtained, but also processed. At the present time the interface of choice seems to be the retinas, and the form of choice seems to be a line graph. Other formats may work just as well. Perhaps for the sight-impaired, the forex data could be translated into sounds. For the time being let us focus on the sight-enabled.

Similar to the technology displayed in the film "lawnmower man", I think that the end result may be a device or program that scrolls through a visual display of historical data. The scale and rate of speed will need to be tweaked for greatest accuracy, however.

I think a possible display would consist of a tiered graph of varying resolutions. Perhaps a double graph, the top layer an hour-by-hour for a 24 hour period, the bottom perhaps a 3 day period.


This project is predicated on the ability of a person to allow their subconscious mind to directly control or influence some external, measurable variable to the end of outputting pattern-analysis results. Making this process as simple and easy as possible may be paramount to getting best results in the shortest amount of time.