How can a derivative estimate a small change in a function’s output?
Quick answer
At a known input a, use the derivative f′(a) as the local rate of change. For a small input change Δx, estimate the output change as f′(a)Δx, so f(a+Δx) is approximately f(a)+f′(a)Δx. This tangent-line estimate is most useful when a+Δx is near a; it is an approximation, not necessarily the exact function value. [S1]
The question, the details, the next step.
I know the derivative tells me a rate of change, but how does that give me an estimate for a function value?
Think of the derivative at a known input as the slope of the tangent line there. Close to that input, the function’s graph can be approximated by that line. At x=a, the tangent-line approximation is L(x)=f(a)+f′(a)(x−a). So if the input moves by Δx, the estimated output change is f′(a)Δx, and the estimated new output is f(a)+f′(a)Δx. [S1]
The key condition is that the new input is near a. You are extending the tangent line a short distance, not claiming that the whole function is a straight line.
Can you show the steps with numbers, including how to tell the estimate from the exact value?
Take f(x)=1/x and estimate f(2.1) using the nearby known input a=2. First, f(2)=1/2=0.5 and f′(x)=−1/x², so f′(2)=−1/4=−0.25. The input change is Δx=2.1−2=0.1.
Now estimate the output change: f′(2)Δx=(−0.25)(0.1)=−0.025. Add that to the known output: 0.5−0.025=0.475. Thus the tangent-line estimate is f(2.1)≈0.475. The exact function value is 1/2.1, about 0.47619, so the estimate is close but not identical. [S1]
The difference is a reminder to label the result as an estimate. The derivative gives the local linear prediction; the exact change is f(a+Δx)−f(a). [S1]
How do I choose the known point and avoid mixing up the change in x with the new x-value?
Choose a point a where you can find both f(a) and f′(a), preferably close to the input you want to estimate. Keep these quantities separate:
- a is the starting input.
- Δx is the change from that input.
- a+Δx is the new input.
For the example, a=2, Δx=0.1, and a+Δx=2.1. In the tangent-line formula, use x−a=Δx. A common slip is to put the new input 2.1 in place of the change 0.1; that would make the tangent-line step much too large. The approximation is intended for a nearby input, as the linearization formula specifies. [S1]
Can I use the estimate as the answer in homework, and what should I write so my method is clear?
Use it when the question asks for an estimate or approximation. If it asks for an exact value, calculate the function value directly when possible; the tangent-line estimate need not equal it. [S1] Show the base point, derivative, input change, and approximate result so the reader can see what you did.
Copyable format:
“Use a = [nearby known input]. Then f(a) = [known output] and f′(a) = [slope]. For the target input, Δx = [target input]−a = [change]. The estimated output is f(a)+f′(a)Δx = [calculation] ≈ [estimate]. This is a local tangent-line estimate, not necessarily the exact value.”
For the example, that becomes: “At a=2, f(2)=0.5 and f′(2)=−0.25. For x=2.1, Δx=0.1, so f(2.1)≈0.5+(−0.25)(0.1)=0.475.”
A practical checklist
Choose a known input a close to the target input.
Calculate f(a) and the derivative value f′(a).
Find Δx by subtracting a from the target input.
Compute f(a)+f′(a)Δx and label it as an estimate.
If needed, compare with the exact function value.
A few more useful answers
Is the estimated output change the same as the exact change?
Not necessarily. The derivative-based estimate is f′(a)Δx; the exact change is f(a+Δx)−f(a). [S1]
Why does the target input need to be near the known input?
The method uses the tangent line to approximate the function near a differentiable point a. It does not establish that the function follows that line far from a. [S1]
What should I do if the problem asks for an exact value?
Evaluate the function at the target input directly when possible. A tangent-line calculation provides an approximation, not a guarantee of exactness. [S1]
Sources
References used for this guide.
- [S1] Linear Approximations and Differentials — Calculus Volume 1, Section 4.2 openstax.org · Checked