I am using feed-forward neural networks for stock prediction, but i am getting problems on training function (pre-define in matlab). Its about input/output size. I set it regarding to function instruction (given in tool), but do not get results.
Please mention what error you got. It is appreciated if you also give some info like dimension of input data and a small sample file.
Error one :
??? error using cat ==>cat
Cat arguments dimensions are not consistent.
when i change data and set it to same dimensions.another Error occur .
Error two:
??? out of memory. ...
Please guide me about selection of data. I want it urgently.. Thanks in advance.
I think first of all you have to be confident about the dimension of your data. cat error occurs when you trying to concatenate data having unequal dimension. for second error I am suggesting you to
1) use single precision in your code
2) check matlab memory. also can use memory command at various place to check explosion in usage. :)
3) check there is no any infinite loop operates within code
4) if using windows turn on data swapping
5) clear variable within your code which has no longer in use
6) at last if using 32 bit machine, its advisable to use 64 bit machine with good RAM
- Badges
- Science topic