Skip to content

Commit

Permalink
Update README.md
Browse files Browse the repository at this point in the history
  • Loading branch information
ahwillia authored Nov 26, 2016
1 parent 8c127bd commit c1db561
Showing 1 changed file with 2 additions and 2 deletions.
4 changes: 2 additions & 2 deletions matlab/README.md
Original file line number Diff line number Diff line change
Expand Up @@ -8,7 +8,7 @@ Then, either download this github repo as a zip file, or clone it from the comma
$ git clone https://github.com/ahwillia/tensor-demo.git
```

Open MATLAB and navigate the folder containing this code and run the script `cp_neuron_demo` to generate some synthetic data and fit a third-order CP decomposition.
Open MATLAB and navigate the folder containing this code and run the script `cp_demo` to generate some synthetic data and fit a third-order CP decomposition.

You can play with various parameters/settings on the synthetic data. As long as noise is low enough, the CP decomposition (fit by alternating least-squares, as reviewed in [Kolda & Bader](http://www.sandia.gov/~tgkolda/pubs/pubfiles/TensorReview.pdf)) should do a pretty good job of estimating the true latent factors:

Expand All @@ -28,7 +28,7 @@ Even in the more difficult scenario on the right, some of the initial guesses yi

### Take Home Message/Recipe

Looking at the internals of `cp_neuron_demo` should give you a basic idea of how to apply CP decomposition to your own data. The essential workflow is embarassingly easy:
Looking at the internals of `cp_demo` should give you a basic idea of how to apply CP decomposition to your own data. The essential workflow is embarassingly easy:

```matlab
data = % your 3-D data (neurons x time x trials)
Expand Down

0 comments on commit c1db561

Please sign in to comment.