Skip to content

CLAUDE.md

Latest commit

 

History

History
72 lines (59 loc) · 3.42 KB

CLAUDE.md

File metadata and controls

72 lines (59 loc) · 3.42 KB
Raw

MachineLearning — Applied Data Analysis and Machine Learning

Teaching and textbook repository of Morten Hjorth-Jensen (University of Oslo / Michigan State University). Course: FYS-STK3155/4155 — Applied Data Analysis and Machine Learning.

Two parallel deliverables

The repository produces two independent products. They cover the same subject matter and should stay topically aligned, but they are separate tracks with separate sources — do not try to auto-generate one from the other.

Track Source Output
Book (PDF) doc/BookML/*.tex — hand-written LaTeX, Springer svmono class doc/BookML/book.pdf
Jupyter-book doc/LectureNotes/*.ipynb + _toc.yml doc/LectureNotes/_build/html/

Directory map

doc/
  BookML/                 The LaTeX book
    book.tex              Root file: preamble, macros, \input{chapterN}
    chapter1.tex          Chapters written directly in LaTeX
    preface.tex acknow.tex acronym.tex dedic.tex
    book.bbl              Checked-in bibliography (no references.bib here yet)
    sp*.bst               Springer bibliography styles
    BookChapters/         Legacy DocOnce sources (chapterN.do.txt) — reference material
    BookPrograms/         Programs accompanying the book (currently empty)
  LectureNotes/           The jupyter-book
    _config.yml _toc.yml  Jupyter-book configuration
    intro.md schedule.md teachers.md textbooks.md
    chapterN.ipynb        Topic chapters (DocOnce-generated, now maintained here)
    statistics.ipynb linalg.ipynb clustering.ipynb chapteroptimization.ipynb
    weekNN.ipynb          Weekly lecture notebooks (semester material)
    exercisesweekNN.ipynb Weekly exercise sets
    projectN.ipynb        Project descriptions
    figures/ figslides/ data/ DataFiles/
    requirements.txt      Pinned build environment (jupyter-book 1.0.4)
    _build/               Generated — never edit
  src/                    DocOnce sources for lectures and topic modules
    weekNN/               weekNN.do.txt, exercisesweekNN.do.txt, make.sh, latex.sh
    <Topic>/              Regression, NeuralNet, CNN, Optimization, ...
  pub/                    Published DocOnce output (html/, ipynb/) — generated
  Programs/               Standalone example programs by topic
  Projects/<year>/        Project sets per year
  HandWrittenNotes/ Articles/ Textbooks/ MathFoundationML/ web/ Admin/

Language and style conventions

  • Python is the primary language; C++ and Fortran appear for performance comparisons and are written in an object-oriented style.
  • Notebooks and Jupyter-books are the preferred way to display and discuss code.
  • Prose is written in the author's voice: direct, pedagogical, "we" rather than "you", equations developed step by step rather than stated.
  • American spelling, LaTeX math throughout ($...$ inline in notebooks).

Skills

Task-specific instructions live in .claude/skills/:

  • ml-book-chapter — write/edit doc/BookML LaTeX chapters, build book.pdf
  • jupyterbook-lectures — notebooks, _toc.yml, jupyter-book builds
  • ml-code — Python/C++/Fortran programs and code cells/listings
  • doconce-ml — DocOnce sources in doc/src and BookML/BookChapters
  • course-weeks — weekly lectures, exercise sets and projects

Never commit or edit

_build/, doc/pub/ (generated), *.aux *.log *.out *.toc *.idx, *~ backup files, .ipynb_checkpoints/, .DS_Store.