Does MATLAB support neural networks for time-series classification problems?
2 views (last 30 days)
Show older comments
I have 300 samples of frequency vs. time data divided into 3 categories (increasing trend, decreasing trend, and normal). Below is a chart of the 3 types of samples clumped into one figure just so you can see clearly what I'm working with.
![](https://www.mathworks.com/matlabcentral/answers/uploaded_files/164963/image.jpeg)
My objective for a project is to see if this data can be run through a network in MATLAB and predict with decent accuracy which of the 3 classes the specific data set falls into. However, I have run into a bit of trouble with this. I may not be noticing something critical, but it appears that toolboxes such as patternnet only work for data that does not need to be in a particular sequence. In addition... other nets in MATLAB aim to predict future activity.. but I am not looking to predict future data, I am looking to classify new data as it is run through. Can MATALB support this kind of mission? Thanks in advance.
1 Comment
Dharmendra Gurve
on 22 Nov 2017
You can take time-frequency representation of the signal (e.g. STFT) as input of the NN.
Answers (1)
Greg Heath
on 9 Jun 2017
The obvious choice is to use the 1st order differences of the smoothed data as the input to the classifier.
Hope this helps.
Thank you for formally accepting my answer
Greg
0 Comments
See Also
Categories
Find more on Sequence and Numeric Feature Data Workflows in Help Center and File Exchange
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!