Stock Market Prediction Program

[seppi85]

Dear Investors

I'm designing a Stock Market Prediction Program for my Final Year Univercity Project.

Computer Jargon aside I hope to use a dozen or so Technical Analysis tools to build up an Neural Network in which it would help the User decide upon what shares to buy/sell etc.

Ive done abit of research and I'm struggling to figure out the most important Technical Analysis Tools.

I guess this is a sort of grey area as everyone has their own prefrences and predicting Stock Market is not real science.(or it would have been done).

All in all I hope active investors could give me tips on what Technical Tools
they use day in day out, and appreciate any feedback you have for me.

Thanks

Seppi

[icjackson]

The (technical/charting) systems that suit me are thus:

one is based on end-of-day prices, using 2 MA lines and just one further indicator

the second, is a scalping method, applied to the Forex market

[goldbug]

Even if you feed 12 technical indicators to the neaural network, they are all different views of the same underlying data (price/volume?). For your neural netwrok, you need to find different data input such as price, volume, sentiment (put/call ratios), and maybe fundamentals such as p/e ratio, earnings, debt, dividend etc although these are hard to find in a timely way.

[spicer]

Hi Sep,

Have you gotten anywhere with this? I would not mind trading notes. I have been doing a lot predictive work with watching the options market. Mainly using end of day data. The system I built back tested over 3 years is about 51% right. Then go through it and use some human thought I can get my success rates up closer to 70%. I have been watching chatter on the web looking to relate chatter as another input to help get better results.

Want to swap ideas?

~Spicer

[seppi85]

Hi Zig

Thank you for your reply.

I'm extracting End of Day Data from a source on the web to my own Database and using Java to communicate with the DB and perform analysis.
so I hope to predict Swing Trading. (and Possibly Long Term Holds)


I guess its highly probable the TI you use for Day Trading is different to Swing and Long-Term.

Could you possibly tell me the differences in strategies between Swing and Long Term and again list TI you use as you kindly did in the previouse post.


Coming from a Engineering background I guess what i meant about Predicting the Stock Market is not an exact science, is that it could be not definitivly proved.
For example I have bought a book "Technical Analysis from A-Z et al Achelis" it lists over 100 TI's.
In my naive point of view, I see things as if it was an exact science. By programming every single TI into my program I should be able to Predict Stock Market with 0 uncertainly.
But that fact that StockMarket depends not just on TI indicators but also World Events, Weather and seasons etc. just makes its impossible to be right 100% of the time?

Anywho Hope that made sense, apperciate any feedback you have for me

Thanks

Seppi

[zigzagman]

Hi Seppi,

What style of trading are you going to do your thesis on? Day Trading, Swing Trading, or Long-Term Holds?
I do all three styles, and I use different Technical Indicators when doing really quick trades as seen in the chart below.

My Day Trading Strategy is based upon Bollinger Bands, the CCI, and Stochastics,
and I use the rest of the indicators to confirm or deny the buy and sell signals I get from my three main ones.

Here are all of the TI's I use: Bollinger Bands, Parabolic SAR, the 5Moving Average (MA) and the 15MA, Chaikin Money Flow (CMF), On Balance Volume (OBV), Accumulation/Distribution (A/D), Commodity Channel Index (CCI), Stochastics (STO), and Volume:



The chart below it shows all of additional the indicators I use for Swing Trading.
I add the Moving Average Convergence/Divergence (MACD), the Average Directional Index (ADX), and the Relative Strength Index (RSI) .

A good website to learn all about these Technical Indicators is:
http://stockcharts.com/school/doku.php?id=chart_school:technical_indicators

Any you are wrong that no one can predict what the market will do next. I do it all the time...

[/b]