AI Blocks › Build and ship software
Prediction feature picker
Give it a data table and a target. Get the best input columns chosen, tested and delivered as a clean table.
Prediction feature picker is a free AI skill for the job "Build and ship software". Give the instructions to an AI assistant that supports the required files and tools, then review its output.
Free · No account · BSD 2 Clause licence · Made by MathWorks · Level: advanced · Thinking: high
Made by MathWorks. See the original on GitHub; we copied it as it is, with its licence.
What it does
Choose the columns that best predict your target, test the choice and receive a clean table ready for training. This is for MATLAB and needs a licence. The process runs in six phases: intake, build a pool of candidate features, select, assess, deliver and report. If your data is images, signals or text, it first routes extraction to a suitable skill. Selection uses several methods and keeps what they agree on, so the result does not rest on one test. It handles one response at a time, either classification or regression, and does not train the final model.
You give it
- Data table
- Target column
- Whether the task is classification or regression
- MATLAB access
You get
- Selected features
- Evidence from several selection methods
- Assessment results
- Clean feature table and report
Use it when
- Your table has 200 columns and you need the useful ones.
- You want proof that your chosen features work.
- Your model is slow and overfitting.
- You must hand a clean feature table to another team.
How to use it
- Add it to your AI once: download it and add the folder to Claude, ChatGPT or Codex as a skill.
- Ask for the job in your own words. Your AI opens the skill when it fits.
- It works best when your AI can run MATLAB with the Statistics and Machine Learning Toolbox.