---
license: mit
dataset_info:
  features:
  - name: L
    dtype: float32
  - name: Fx
    dtype: float32
  - name: Fy
    dtype: float32
  - name: xP
    dtype: float32
  - name: Fy2
    dtype: float32
  - name: xQ
    dtype: float32
  - name: Ax
    dtype: float32
  - name: Ay
    dtype: float32
  - name: By
    dtype: float32
  - name: score_minmax
    dtype: float32
  - name: score_euclid
    dtype: float32
  - name: score_balance
    dtype: float32
  - name: winner
    dtype:
      class_label:
        names:
          '0': minmax
          '1': euclid
          '2': balance
  splits:
  - name: original
    num_bytes: 1680
    num_examples: 30
  - name: augmented
    num_bytes: 168000
    num_examples: 3000
  download_size: 201706
  dataset_size: 169680
configs:
- config_name: default
  data_files:
  - split: original
    path: data/original-*
  - split: augmented
    path: data/augmented-*
---

# Beam Tabular HW1

This dataset was created for CMU 24-679 (AI/ML for Engineers) HW1.  
It extends Project 0 (beam statics simulator) into a tabular dataset suitable for ML practice.

## Splits
- **original**: 30 manually specified load cases
- **augmented**: 3000 synthetic cases generated by physics-based random sampling

## Features
- `L`: beam length [m]  
- `Fx`, `Fy`: applied force components [N]  
- `xP`: location of primary load [m]  
- `Fy2`, `xQ`: optional surprise vertical load and its location [m]  
- `Ax`, `Ay`, `By`: reaction forces at supports [N]  
- `score_minmax`: max reaction criterion  
- `score_euclid`: Euclidean norm of reactions  
- `score_balance`: balance score between supports  
- `winner`: best objective (0=minmax, 1=euclid, 2=balance)

## Augmentation Method
We used **Monte Carlo random sampling** of beam parameters within realistic ranges, then recomputed support reactions and optimizer scores. This approach ensures every augmented sample obeys statics, while covering a wider design space than the 30 original cases.

## License
MIT — released for educational use.

## AI Tool Disclosure
ChatGPT (OpenAI, GPT-5) was used for:  
- Code support  
- Documentation writing  

All diagrams, dataset generation, and labeling logic were created and validated by the author.