How to ask MATLAB to predict the next value in a graph

I have a linear graph with the cost of advertising on a platform over time, versus the date. I am looking to use Matlab to estimate the next point on the graph.

I have looked at YPred, but I am unsure about how to format my data so that it understands what it needs to make the prediction. I am also unclear as to what type of model I would be specifying.

ypred = predict(mdl,DATASET)

is what I am using, but it doesn't make any sense.

Any suggestions or advice would be appreciated, I have exhausted the MATLAB resources sadly.

Answer 1

[Image Analyst]

Use polyfit() and polyval():

coefficients = polyfit(t, cost, 1); % Do a linear fit
predictedCost = polyval(coefficients, t); % Get value of linear fit at time t.

[Helena Leathers]

Thank you! This worked, but it returned a whole bunch of data that is linear. The data it returned started from the lowest point of my original graph and stepped up to the highest.

How could I ask MatLab to recalculate the next value in addition to the previous one it found? Applicability wise, the first data point is interesting but a linear pattern isn't like reality. Is there a way to get the program to consider this?

Thanks again, I spent hours trying to figure this out on my own. I am eternally grateful!

Answer 2

[Cris LaPierre]

You probably want to extrapolate. See this answer.

You could also consider fitting the data using polyfit, then calculating the y value for a given x value using polyval. The this example in the documentation.

[Helena Leathers]

Hey yes this option also worked and was super simple. Thanks!