Visualizing Taylor Series Convergence
Taylor series are a powerful tool in mathematics, allowing us to approximate complex functions using infinite sums of polynomials. But what does convergence really *mean*? And how do you build intuition for it?
The Taylor Series
For a function f(x) expanded around point a = 0 (Maclaurin series):
``
f(x) = Σₙ₌₀^∞ f⁽ⁿ⁾(0) / n! · xⁿ
The more terms you include, the better the approximation — at least within the radius of convergence.
## Classic Examples
sin(x)
sin(x) ≈ x − x³/3! + x⁵/5! − x⁷/7! + ...
Converges for all real x. Needs more terms for large |x|.
### cos(x)
cos(x) ≈ 1 − x²/2! + x⁴/4! − x⁶/6! + ...
Also converges for all real x. Note sin²+cos²=1 holds exactly when enough terms are used.
### eˣ
eˣ ≈ 1 + x + x²/2! + x³/3! + ...
Converges for all real x, and converges quickly — the factorial denominator grows fast.
### ln(1+x)
ln(1+x) ≈ x − x²/2 + x³/3 − x⁴/4 + ...
Converges only for |x| ≤ 1 (with x ≠ −1). This is the alternating harmonic series at x = 1 and converges slowly.
## Key Concepts
### Radius of Convergence
The set of x values for which a series converges. For polynomials like eˣ, sin, and cos, this is all of ℝ. For ln(1+x), it is [−1, 1).
### Rate of Convergence
Different series converge at very different speeds. The eˣ series is fast because the factorial denominator shrinks each term rapidly. The ln(1+x) series at x=1 converges as slowly as 1/n, requiring hundreds of terms for high accuracy.
### Remainder Term
The error after n terms is bounded by the (n+1)th term in the series — useful for estimating how many terms you need.
## Try It in the Lab
The Taylor Series Lab in IngenioLens shows both the exact function and the n-term approximation side by side. Try:
- Increasing the term count from 1 to 15 for sin(x) — watch the approximation spread outward
- Comparing eˣ vs ln(1+x) at n=10 — notice how eˣ is already nearly perfect while ln still has visible error at |x| near 1
- Setting n=1 for any function — you recover the linear (tangent line) approximation
For a function f(x) expanded around point a = 0 (Maclaurin series):
``
f(x) = Σₙ₌₀^∞ f⁽ⁿ⁾(0) / n! · xⁿ
The more terms you include, the better the approximation — at least within the radius of convergence.
## Classic Examples
sin(x)
sin(x) ≈ x − x³/3! + x⁵/5! − x⁷/7! + ...
Converges for all real x. Needs more terms for large |x|.
### cos(x)
cos(x) ≈ 1 − x²/2! + x⁴/4! − x⁶/6! + ...
Also converges for all real x. Note sin²+cos²=1 holds exactly when enough terms are used.
### eˣ
eˣ ≈ 1 + x + x²/2! + x³/3! + ...
Converges for all real x, and converges quickly — the factorial denominator grows fast.
### ln(1+x)
ln(1+x) ≈ x − x²/2 + x³/3 − x⁴/4 + ...
Converges only for |x| ≤ 1 (with x ≠ −1). This is the alternating harmonic series at x = 1 and converges slowly.
## Key Concepts
### Radius of Convergence
The set of x values for which a series converges. For polynomials like eˣ, sin, and cos, this is all of ℝ. For ln(1+x), it is [−1, 1).
### Rate of Convergence
Different series converge at very different speeds. The eˣ series is fast because the factorial denominator shrinks each term rapidly. The ln(1+x) series at x=1 converges as slowly as 1/n, requiring hundreds of terms for high accuracy.
### Remainder Term
The error after n terms is bounded by the (n+1)th term in the series — useful for estimating how many terms you need.
## Try It in the Lab
The Taylor Series Lab in IngenioLens shows both the exact function and the n-term approximation side by side. Try:
- Increasing the term count from 1 to 15 for sin(x) — watch the approximation spread outward
- Comparing eˣ vs ln(1+x) at n=10 — notice how eˣ is already nearly perfect while ln still has visible error at |x| near 1
- Setting n=1 for any function — you recover the linear (tangent line) approximation