Artificial Neural Network Long-Range Forecasting

I am working on a project to predict stock prices using ANN. I trained the system using previous data for 7 years, and it works great to predict data in one day. Now I want to predict the price of shares over the next seven days.

My idea is to predict the data of day 2 using the data of day 1, day 3 using the predicted data of day 2 and day 1, etc. But it does not work properly.

I trained ANN to predict the closing price using the opening price, maximum and minimum daily prices.

What is the idea of ​​predicting data from the next seven days?

+5
source share
2 answers

, ANN 1 , , , ! LOL

. - , - . , ( ), . , , - :

let min = -0.5
let max = +0.5
let bias = 0.01
let random = rand(min, max)
y[i] = y[i-1] + random + bias

( -0.01 0.01), , LOT, . , , , . , 55% ? , ...

, , . . , .

: ANN, 1 ? , , (, , , ..), , , 1- .

Edit:

-, , ? mikera, ​​, .

1000 , 800 ANN. . (, ) (% ) . 1 . ANN 1 , .

200 , (, )? 10% ? ANN , ? , ...

, .

+3

, - :

  • , , . ? , ANN 20% 80%, .
  • ? , , , , ( ). , 99% + R-, ....

, , , . .. ( ), , ( , , ..).

, / ( ). , .

+2

All Articles