Neural Network Approach for Market Prediction : Inputs

Dileepa

Dileepa

Senior Manager - Equity Analytics

Posts : 346
Join date : 2011-03-21

slstock wrote:
Are you training on overall CSE data a set or just on a single company ?

If one company just an example:
To test you could add
a) +5% increase on a share split annoucement day
b) -4% drop on 2 days afterwards
b) + 5% on the day the shares have been split
c) - 4% on day follow the split

Again these numbers can vary but giving you an idea.

For Dividend you could have 2% share price increase as an example on the annoucement date , then a 1% drop the next few days etc..You get the idea

You could test and see.

For the time being thought of concentrating on BFI sector (to be specifically NTB).
Attached file is a screen shot of sample test data.
(sorry about the colors, had to make it mono due to upload limit)
And thanks for the suggestion, will post some more results again Smile

yoyo

yoyo

Manager - Equity Analytics

Posts : 202
Join date : 2011-02-22

Good work Dileepa. + Rep from me. Keep up ur good work.

mono

mono

Vice President - Equity Analytics

Posts : 1167
Join date : 2011-01-21

keep the H, L, O, C, volume & ASI + maybe MPI

Everything else is useless.

wikum100

wikum100

Senior Manager - Equity Analytics

Posts : 303
Join date : 2011-01-28

good work deelipa ,eventhough i coudnt understand much of it it seems your are trying to do a marvelous work it fully succes...wish you all very best +rep from me to your hard work Very Happy

wikum100

wikum100

Senior Manager - Equity Analytics

Posts : 303
Join date : 2011-01-28

Chinwi wrote:

Quibit wrote:I simply do not understand the concept and the charts. What are you trying to do? I only can see a JPG file, No program or a software. Can any one of the expert help me to understand this!

Neural Networks are basics of Artificial Intelligence. The programmers try to mimic actual world relations as far as possible in the model developed inside the computer.

More relations and interactions you can think and include in the system will bring it closer to functioning of our brain and produce better results. I think you can write the program to do self adjustments or fine tune when the system is fed with more and more actual data comparing its own predictions.

In Dileepa's program Green line is the predicted out put while
Magenta line represents the actual data of NTB.
With the time, he can fine tune the program to make the two lines closer and closer.

thanks for your nice explanation ..now i got understood +rep from me

Genting

Genting

Senior Manager - Equity Analytics

Posts : 349
Equity Stars : 5311
Reputation : 46
Join date : 2011-02-21

Dileepa

Nice work!

Can I know, after training, what are the weights you ended up for those inputs?

I also tried design similar thing but realizes that this stock market is 99% irrational!

wmdcf

wmdcf

Manager - Equity Analytics

Posts : 126
Equity Stars : 6062
Reputation : 17
Join date : 2011-01-29

Good work!
I'm also did some kind of same thing(few month back), but end up with big losses. Smile

I think we need to find some other SE other than CSE to test this.

JennyFunny

JennyFunny

Senior Manager - Equity Analytics

Posts : 452
Equity Stars : 5254
Reputation : 12
Join date : 2011-05-20
Age : 33
Location : Colombo

could some explain how you do this?

Dileepa

Dileepa

Senior Manager - Equity Analytics

Posts : 346
Equity Stars : 6259
Reputation : 44
Join date : 2011-03-21
Location : Tokyo

mono wrote:keep the H, L, O, C, volume & ASI + maybe MPI

Everything else is useless.

Is the opening val equal to the previous day closing?
if both are same, closing only will be enough I think Exclamation

Dileepa

Dileepa

Senior Manager - Equity Analytics

Posts : 346
Equity Stars : 6259
Reputation : 44
Join date : 2011-03-21
Location : Tokyo

Genting wrote:Dileepa

Nice work!

Can I know, after training, what are the weights you ended up for those inputs?

I also tried design similar thing but realizes that this stock market is 99% irrational!

Actually weights are not that important, it is the network design which matters I think.
I used a feed forward back propagation network with one hidden layer containing 13 neurons.

Dileepa

Dileepa

Senior Manager - Equity Analytics

Posts : 346
Equity Stars : 6259
Reputation : 44
Join date : 2011-03-21
Location : Tokyo

wmdcf wrote:Good work!
I'm also did some kind of same thing(few month back), but end up with big losses.
I think we need to find some other SE other than CSE to test this.

sorry about your losses, training the network is equally important as the design.
can you tell about the data and training process you followed.
At first,
I trained the system with 200 data records for 10,000 cycles, but the accuracy wasn't that great.
Then I used 400 records and trained only for 3000 cycles, but it reached 80%++ level of accuracy.
Random behavior is a problem, but NN can cope it with some fine tuning.
(by the way you should have posted your findings at the first place,
so that others will not have to travel the same path again)

samueldaniel7

samueldaniel7

Manager - Equity Analytics

Posts : 128
Equity Stars : 5021
Reputation : 33
Join date : 2011-03-21

can we have a copy of it

Dileepa

Dileepa

Senior Manager - Equity Analytics

Posts : 346
Equity Stars : 6259
Reputation : 44
Join date : 2011-03-21
Location : Tokyo

samueldaniel7 wrote:can we have a copy of it

All the data I'm using was collected from central bank and stocksyn.(available online)
I will receive that CSE data CD soon. I may be able to upload the content once I received that.

samueldaniel7

samueldaniel7

Manager - Equity Analytics

Posts : 128
Equity Stars : 5021
Reputation : 33
Join date : 2011-03-21

sorry i was asking about the neural network script, did you make it or did you download it from anywhere,
if you dont mind can we have a copy

Dileepa

Dileepa

Senior Manager - Equity Analytics

Posts : 346
Equity Stars : 6259
Reputation : 44
Join date : 2011-03-21
Location : Tokyo

samueldaniel7 wrote:sorry i was asking about the neural network script, did you make it or did you download it from anywhere,
if you dont mind can we have a copy

I'm not familiar with java much, so there won`t be a script.
Still I`m trying to get the design correct.
Currently modeling it using NN design software (NN designer etc.)
Will code it using C/C++ later on, but first I want to make sure it worth a shot.
If you are interested, below design is the one I'm trying to fine tune.
(though it is really dumb for the moment rendeer

)

Neural Network Approach for Market Prediction : Inputs - Page 2 Networ10

samueldaniel7

samueldaniel7

Manager - Equity Analytics

Posts : 128
Equity Stars : 5021
Reputation : 33
Join date : 2011-03-21

Thanks a lot

Ben Graham

Ben Graham

Senior Equity Analytic

Posts : 76
Equity Stars : 6023
Reputation : 12
Join date : 2011-02-07

nice work dileepa. + rep from me too.

san1985

san1985

Senior Manager - Equity Analytics

Posts : 400
Equity Stars : 5323
Reputation : 16
Join date : 2011-01-30

Great work buddy....Keep up ur good work..

But i dont know to what extent this would work in CSE...

If u can add manipulation as a variable this would certainly work......

Sponsored content

Name*:
Phone number:
Email address*:
Comments*:
* - required fields	Learn

Name*:
Phone number:
Email address*:
Comments*:
* - required fields	Learn

Latest topics

Neural Network Approach for Market Prediction : Inputs

26 Re: Neural Network Approach for Market Prediction : Inputs Sun May 29, 2011 7:34 pm

Dileepa

27 Re: Neural Network Approach for Market Prediction : Inputs Sun May 29, 2011 9:22 pm

yoyo

28 Re: Neural Network Approach for Market Prediction : Inputs Sun May 29, 2011 10:02 pm

mono

29 Re: Neural Network Approach for Market Prediction : Inputs Mon May 30, 2011 12:09 am

wikum100

30 Re: Neural Network Approach for Market Prediction : Inputs Mon May 30, 2011 12:12 am

wikum100

31 Re: Neural Network Approach for Market Prediction : Inputs Mon May 30, 2011 8:27 am

Genting

32 Re: Neural Network Approach for Market Prediction : Inputs Mon May 30, 2011 9:03 am

wmdcf

33 Re: Neural Network Approach for Market Prediction : Inputs Mon May 30, 2011 11:53 am

JennyFunny

34 Re: Neural Network Approach for Market Prediction : Inputs Mon May 30, 2011 7:07 pm

Dileepa

35 Re: Neural Network Approach for Market Prediction : Inputs Mon May 30, 2011 7:18 pm

Dileepa

36 Re: Neural Network Approach for Market Prediction : Inputs Mon May 30, 2011 7:40 pm

Dileepa

37 Re: Neural Network Approach for Market Prediction : Inputs Mon Jun 06, 2011 5:46 pm

samueldaniel7

38 Re: Neural Network Approach for Market Prediction : Inputs Sun Jun 12, 2011 7:01 pm

Dileepa

39 Re: Neural Network Approach for Market Prediction : Inputs Sun Jun 12, 2011 8:03 pm

samueldaniel7

40 Re: Neural Network Approach for Market Prediction : Inputs Mon Jun 13, 2011 6:29 pm

Dileepa

41 Re: Neural Network Approach for Market Prediction : Inputs Mon Jun 13, 2011 10:45 pm

samueldaniel7

42 Re: Neural Network Approach for Market Prediction : Inputs Tue Aug 16, 2011 11:18 pm

Ben Graham

43 Re: Neural Network Approach for Market Prediction : Inputs Tue Aug 30, 2011 8:54 pm

san1985

44 Re: Neural Network Approach for Market Prediction : Inputs

Sponsored content

New topics