Skip to content

6. Signals

Julio Toboso edited this page Sep 9, 2026 · 1 revision

A signal is a number linked to something. When a number is analyzed without context, it is part of maths. When it relates to anything else, it is a signal.

Usually signals link a measurement to a sample. These measurements may provide more than one number, so a signal may contain more than one value per sample, and usually a single measurement doesn't provide context or utility, so we make multiple measurements. We need to organize these measurements by a parameter we call the Index.

In many applications the index is time, or location, or a combination of time and location, and a measurement records an event for further analysis. But the index can be anything that allows differentiation of the measurements.

Mathematics and signals share a common essence. In nature and behavior. Signals tend to apply mathematics into the world.

Data

Legend:

  • 🔰 Start here
  • 🏵️ Important Insight
  • ⚙️ Technical and applications
  • 💻 Some Code
  • ⚜️ Advanced Topics

Start

01 · Ω — What could happen

02 · X, 𝒳 — What we observe

03 · p(x) — What we expected

04 · I(x), H(X) — What an observation tells us

05 · Data — What is consistent

06 · Statistics — What we can learn.

07 · Databases - How to store data

08 · μ(x) — Fuzzy - When data is flexible

  • 08.1 Fuzzy Logic (planned)
  • 08.2 Fuzzy Numbers (planned)
  • 08.3 Possibility vs Probability (planned)

09 · x(t) — Signals and systems - Where patterns live

  • 09.1 Signals (stub)
  • 09.2 Sampling and Aliasing (planned)
  • 09.3 Convolution (planned)
  • 09.4 Frequency (planned)

10 · f_θ — Modelling and linear algebra

  • 10.1 🔰 Vectors and Matrices (planned)
  • 10.2 Mathematical Modelling (planned)

11 · min L(θ) — Optimisation

  • 11.1 🔰 Derivatives (planned)
  • 11.2 Convex Optimisation (planned)

12 · Learning from data

13 · Cause, design, capstone

  • 13.1 Cause and Effect (planned)
  • 13.2 Capstone Guide (planned)

Appendix

Clone this wiki locally