"Data has less than 10 timestamps" when using 3000+ data points with Neural Network NARX tool?
3 次查看(过去 30 天)
显示 更早的评论
I'm trying to use stock data from Yahoo in the Neural Network Time Series tool (ntstool). I have the original spreadsheet attached titled "Stock Prices Yahoo CSV". I'm trying to predict closing prices, so I created a new document with only that column's information (attached as "Sheet 1-Table 1-1"). I imported this data into the wizard, selected NARX, and on the Select Data screen I imported that data.
Now for the problem: I'm getting an error saying that the data has less than 10 timestamps. I imported over 3000 data points, which is clearly greater than 10. There must be something I'm doing wrong, or that I'm interpreting wrong, but I can't seem to figure it out. Below is a screenshot of the problem.
0 个评论
采纳的回答
Walter Roberson
2015-10-4
I suspect you should be using matrix column rather than cell column.
I notice you seem to be using the same input variable "data" for both timeseries and target; I'm not sure that is correct. It might be, but it looks suspicious.
2 个评论
Walter Roberson
2015-10-5
Have another look at the screenshot, on the right hand side. Notice that it describes "data" as being 1 sample of 3473 elements, so it thought you only had one timepoint with lots of components. When you select Matrix Column it knows to treat it as 3473 time points of 1 element.
更多回答(1 个)
Greg Heath
2015-10-8
The input and target should be a 1 dimensional row cell sequence. The conversion for matrix timeseries to cell timeseries is given by
DATA = con2seq(data');
Hope this helps.
Thank you for formally accepting my answer
Greg
0 个评论
另请参阅
类别
在 Help Center 和 File Exchange 中查找有关 Sequence and Numeric Feature Data Workflows 的更多信息
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!