Course contentsShow
← Machine Learning and Deep Learning
Lesson 38 of 3,5381. Mathematical Foundations for Machine LearningFree lesson

Derivatives of Trigonometric Functions

Computing derivatives of sine, cosine, and tangent functions used in signal processing.

Derivatives of Trigonometric Functions

What You'll Learn

You'll discover how to compute derivatives of sine, cosine, and tangent functions—essential tools for analyzing periodic patterns in signal processing applications.

Why Trigonometric Derivatives Matter

In signal processing (audio, images, time-series data), machine learning models often encounter periodic patterns. Trigonometric functions naturally describe these wave-like behaviors. To optimize models that process such signals, we need their derivatives.

The Core Derivatives

Here are the three fundamental trigonometric derivatives you must memorize:

Sine Function

The derivative of sin(x) is cos(x):

d/dx[sin(x)] = cos(x)

Think of it this way: when sine is climbing steeply (positive slope), cosine is near its peak (positive value). When sine flattens at its peak, cosine crosses zero.

Cosine Function

The derivative of cos(x) is -sin(x):

d/dx[cos(x)] = -sin(x)

Notice the negative sign! Cosine starts at its maximum and initially decreases, hence the negative derivative.

Tangent Function

The derivative of tan(x) is sec²(x) or equivalently 1 + tan²(x):

d/dx[tan(x)] = sec²(x) = 1/cos²(x)

Applying the Chain Rule

When these functions have compositions (like sin(3x) or cos(x²)), use the chain rule you learned earlier:

d/dx[sin(3x)] = cos(3x) · 3 = 3cos(3x)
d/dx[cos(x²)] = -sin(x²) · 2x = -2x·sin(x²)

Signal Processing Connection

In audio or image filters, gradients involving trigonometric functions help the model learn which frequencies matter most—critical for tasks like noise reduction or feature extraction.

Key Takeaway: The derivatives of sin(x), cos(x), and tan(x) are cos(x), -sin(x), and sec²(x) respectively, enabling gradient-based optimization in models that process periodic signals.