Plotting simplenarx_dataset

I tried to plot the simplenarx_dataset but it is coming up blank (i get a graph with x and y axis but no data in it). Can anyone help me?

load simplenarx_dataset;
data = simplenarx_dataset;
figure; plot(data{1});
hold on;
plot(data{2});
xlabel('Time');
ylabel('Value');
legend('Input', 'Output');

[Dyuman Joshi]

You will need to provide the file simplenarx_dataset for us to give any working suggestion/solution.

Please attach the file via the paperclip button.

[Gibbov]

It is built-in to MATLAB as far as I know. NARX dataset is a 1x100 cell array of scalar values, which represent an 100 timestep time-series in addition to a 1x100 cell array of scalar values representing a 100 timestep time-series for the prediction.

If you type "load simplenarx_dataset;" into MATLAB it creates something, without the need for a file from me.

[Dyuman Joshi]

I see, I was not aware of this particular dataset.

Answer 1

[Voss]

simplenarx_dataset.mat is a mat-file; simplenarx_dataset.m is a function m-file.

When you load simplenarx_dataset, you load the variables contained in the mat-file into the workspace. These variables are simplenarxInputs and simplenarxTargets.

load simplenarx_dataset
whos

  Name                   Size             Bytes  Class    Attributes

  cmdout                 1x33                66  char               
  simplenarxInputs       1x100            11200  cell               
  simplenarxTargets      1x100            11200  cell               

When you call the function simplenarx_dataset, you get up to two outputs, which are the same two variables contained in the mat-file.

[inputs,targets] = simplenarx_dataset();
whos

  Name                   Size             Bytes  Class    Attributes

  cmdout                 1x33                66  char               
  inputs                 1x100            11200  cell               
  simplenarxInputs       1x100            11200  cell               
  simplenarxTargets      1x100            11200  cell               
  targets                1x100            11200  cell               

isequal(inputs,simplenarxInputs) % the first output from simplenarx_dataset() is simplenarxInputs from the mat-file

ans = logical
   1

isequal(targets,simplenarxTargets) % the second output from simplenarx_dataset() is simplenarxTargets from the mat-file

ans = logical
   1

You can inspect the contents of the m-file to see that it is in fact loading the mat-file and returning the requested variables (scroll down to the end):

type simplenarx_dataset.m

function [inputs,targets] = simplenarx_dataset
%SIMPLENARX_DATASET Simple time-series prediction dataset
%
% Input-output time series problems consist of predicting the next value
% of one time-series given another time-series. Past values of both series
% (for best accuracy), or only one of the series (for a simpler system)
% may be used to predict the target series.
%
% This dataset can be used to demonstrate how a neural network can be
% trained to make predictions.
%
% LOAD simplenarx_dataset.MAT loads these two variables:
%
%   simplenarxInputs - a 1x100 cell array of scalar values representing
%   a 100 timestep time-series.
%
%   simplenarxTargets - a 1x100 cell array of scalar values representing
%   a 100 timestep time-series to be predicted.
%
% [X,T] = simplenarx_dataset loads the inputs and targets into
% variables of your own choosing.
%
% You can import this example data set from within the Neural Net Time
% Series app. Open the app by calling ntstool.
%
% Here is how to design a neural network that predicts the target series
% from past values of inputs and targets. See narxnet for more details.
%
%   [X,T] = simplenarx_dataset;
%   net = narxnet(1:2,1:2,10);
%   [Xs,Xi,Ai,Ts] = preparets(net,X,{},T);
%   net = train(net,Xs,Ts,Xi,Ai);
%   view(net)
%   Y = net(Xs,Xi,Ai)
%   plotresponse(Ts,Y)
%
% Here is how to design a neural network that predicts the target series
% from only using past values of inputs. See timedelaynet for details.
%   
%   net = timedelaynet(1:2,10);
%   [Xs,Xi,Ai,Ts] = preparets(net,X,T);
%   net = train(net,Xs,Ts,Xi,Ai);
%
% Here is how to design a neural network that predicts the targets series
% only using past values of the target series. See narnet for details.
%
%   net = narnet(1:2,10);
%   [Xs,Xi,Ai,Ts] = preparets(net,{},{},T);
%   net = train(net,X,T,Xi,Ai);
%
% See also NTSTOOL, NARXNET, TIMEDELAYNET, NARNET, PREPARETS, NNDATASETS.

% Copyright 2010-2021 The MathWorks, Inc.

load simplenarx_dataset
inputs = simplenarxInputs;
targets = simplenarxTargets;

That's all to say that, in your code, the load is unnecessary because you don't do anything with the loaded variables simplenarxInputs and simplenarxTargets. Instead you are calling simplenarx_dataset.m on the next line, with one output, which makes your variable data have the value of simplenarxInputs from the mat-file, which is the first output of the function simplenarx_dataset.

Now for the question you asked: why do plot(data{1}) and plot(data{2}) not produce any visible result?

That's because data{1} and data{2} are each scalars, i.e., the cells in positions 1 and 2 (in fact in all positions) of data each contain only a single value, so you are plotting only single points. You cannot see a single point unless you use a data marker.

figure;
data = simplenarx_dataset();
plot(data{1},'o'); % marker 'o'
hold on;
plot(data{2},'x'); % marker 'x'
xlabel('Time');
ylabel('Value');
legend('Input', 'Output');

Maybe you want to get both outputs from simplenarx_dataset(), and plot them against each other.

figure
[data,targets] = simplenarx_dataset();
data = [data{:}];
targets = [targets{:}];
plot(data,targets,'.')

[Gibbov]

Thanks. I need to plot all of the data on a graph, like the second example, but I am unsure how I would then split the data into two parts for training and testing. Eg. 80% for training, 20% for testing.

[Voss]

https://www.mathworks.com/matlabcentral/answers/377839-split-training-data-and-testing-data

[Gibbov]

Also, how would i connect the dots (with a line between them ) on the graph to create a line between each '.'?

[Voss]

plot(data,targets,'.-')

Answer 2

[Dyuman Joshi]

simplenarx_dataset is a 1x100 cell array, with each cell element containing a numerica scalar.

load simplenarx_dataset;
data = simplenarx_dataset

data = 1×100 cell array
  Columns 1 through 14

    {[0.3112]}    {[0.5285]}    {[0.1656]}    {[0.6020]}    {[0.2630]}    {[0.6541]}    {[0.6892]}    {[0.7482]}    {[0.4505]}    {[0.0838]}    {[0.2290]}    {[0.9133]}    {[0.1524]}    {[0.8258]}

  Columns 15 through 28

    {[0.5383]}    {[0.9961]}    {[0.0782]}    {[0.4427]}    {[0.1067]}    {[0.9619]}    {[0.0046]}    {[0.7749]}    {[0.8173]}    {[0.8687]}    {[0.0844]}    {[0.3998]}    {[0.2599]}    {[0.8001]}

  Columns 29 through 42

    {[0.4314]}    {[0.9106]}    {[0.1818]}    {[0.2638]}    {[0.1455]}    {[0.1361]}    {[0.8693]}    {[0.5797]}    {[0.5499]}    {[0.1450]}    {[0.8530]}    {[0.6221]}    {[0.3510]}    {[0.5132]}

  Columns 43 through 56

    {[0.4018]}    {[0.0760]}    {[0.2399]}    {[0.1233]}    {[0.1839]}    {[0.2400]}    {[0.4173]}    {[0.0497]}    {[0.9027]}    {[0.9448]}    {[0.4909]}    {[0.4893]}    {[0.3377]}    {[0.9001]}

  Columns 57 through 70

    {[0.3692]}    {[0.1112]}    {[0.7803]}    {[0.3897]}    {[0.2417]}    {[0.4039]}    {[0.0965]}    {[0.1320]}    {[0.9421]}    {[0.9561]}    {[0.5752]}    {[0.0598]}    {[0.2348]}    {[0.3532]}

  Columns 71 through 84

    {[0.8212]}    {[0.0154]}    {[0.0430]}    {[0.1690]}    {[0.6491]}    {[0.7317]}    {[0.6477]}    {[0.4509]}    {[0.5470]}    {[0.2963]}    {[0.7447]}    {[0.1890]}    {[0.6868]}    {[0.1835]}

  Columns 85 through 98

    {[0.3685]}    {[0.6256]}    {[0.7802]}    {[0.0811]}    {[0.9294]}    {[0.7757]}    {[0.4868]}    {[0.4359]}    {[0.4468]}    {[0.3063]}    {[0.5085]}    {[0.5108]}    {[0.8176]}    {[0.7948]}

  Columns 99 through 100

    {[0.6443]}    {[0.3786]}

When plotting a numeric scalar via plot, you will need to specify a marker to show the point on the figure -

figure;
plot(data{1},'r*');
hold on;
plot(data{2},'bo');
xlabel('Time');
ylabel('Value');
legend('Input', 'Output');

You can also use scatter to show scalar points on figure, which does not require markers input for that.

[Gibbov]

Thanks. I need to plot all of the data at the same time, not just one bit of data.