I keep getting error using plot not enough input arguments.

I keep getting error when I try to plot the following code and I don't really know what to do.

figure
plot(recall,precision)
xlabel("Recall")
ylabel("Precision")
grid on
title(sprintf("Average Precision = %.2f",ap))

Error using plot

Not enough input arguments.

[Torsten]

And how are the arrays for recall and prescision defined in your code ?

At the moment, both plot components are empty - thus nothing to plot.

[Adrian Kleffler]

[ap,recall,precision] = evaluateDetectionPrecision(detectionResults,testData);

like this

[Adrian Kleffler]

i improved the code to something like this :

recallv = cell2mat(recall);
precisionv = cell2mat(precision);
figure
plot(recallv,precisionv)
xlabel("Recall")
ylabel("Precision")
grid on
title(sprintf("Average Precision = %.2f",ap))

the graph is shown but it looks bad, something else is wrong also.

Answer 1

[Dyuman Joshi]

You need to sort your x data.

recallv = cell2mat(recall);
precisionv = cell2mat(precision);
[r,index] = sort(recallv);
p = precisionv(index);
figure
plot(r,p)
xlabel("Recall")
ylabel("Precision")
grid on
title(sprintf("Average Precision = %.2f",ap))

[Adrian Kleffler]

thanks, it looks like this helped, but i have 6 classes in my detector... it looks like there is average precision for all the classes in one graph ... how to improve it ?

[Dyuman Joshi]

Could you please attach your data?

[Adrian Kleffler]

which data do you mean ?

[Dyuman Joshi]

The arrays - recall and precision.

Also, could you elaborate on this statement?

"it looks like there is average precision for all the classes in one graph ... how to improve it ? "

[Adrian Kleffler]

recall :

0

[0;0.166666666666667;0.333333333333333;0.500000000000000;0.666666666666667;0.833333333333333]

20x1 double

[0;0.125000000000000;0.250000000000000;0.375000000000000;0.500000000000000;0.625000000000000;0.750000000000000;0.875000000000000;1]

12x1 double

[0;0.111111111111111;0.222222222222222;0.333333333333333;0.444444444444444;0.555555555555556;0.666666666666667;0.777777777777778;0.888888888888889;1]

precision :

1

[1;1;1;1;1;1]

20x1 double

[1;1;1;1;1;1;1;1;1]

12x1 double

[1;1;1;1;1;1;1;1;1;1]

i want to say that i have 6 classes in my project, so in that graph (on the top of the graph, ap = 0,83, ap = 0,29...) i posted, there are average precisions for 6 classes in my opinion ... i don´t know if its possible to combine them to only one average precision or what to do with them... on matlab page there is one example where the graph looks like this (but they had only one class to detect) :

I hope you understand me, and i am so thankful for your answers.

[Adrian Kleffler]

maybe it will be easier to make one graph for each class, but i don´t know how to do it ...