mateusfkschenato/arcanum

How mateusfkschenato/arcanum works

Arcanum is a high-load prediction project for electric-arc-furnace scrap baskets, with a Streamlit race simulator and a batch train/predict pipeline. The documented analytical workflow runs from raw EVCOMX data through cleaning, feature engineering, model selection, evaluation, validation, and interpretation; the operational workflow lets an operator assemble a race and inspect the model’s risk result. Edges reflect README-documented stages or sampled imports/calls; some analytical stage internals and batch wiring are unsampled, so their links are documented rather than inferred from names.

Main entry points: protocolo.py Model protocol · selecao_algoritmos.py Model selection · motor.py Simulation engine · app.py Race builder

Components 20

Model development

Clean and explore src/EDA/EDA.py
Feature selection src/feature_selection/feature_selection.py
Model selection src/selecao_algoritmos/selecao_algoritmos.py
Model analysis src/analise_modelos/analise_modelos.py
Verification src/verificacao_validacao/verificacao_validacao.py
Model interpretation src/interpretacao_modelos/interpretacao_modelos.py
Operational scenarios src/otimizacao_minutos/cenarios_operacionais.py

Batch prediction

Train model pipeline/treinar.py
Race features pipeline/features_corrida.py
Batch prediction pipeline/prever.py

Race simulator

Race builder app_streamlit/app.py
Simulation engine app_streamlit/motor.py
Basket display app_streamlit/desenho.py
Simulator data app_streamlit/dados_app

Shared model logic

Model protocol comum/protocolo.py
Recipe recalculation comum/receita.py

Other components

Operator
EVCOMX raw data original_dataset
Race files previsao/entrada
Prediction files previsao/saida

Connections 24

EVCOMX raw data → Clean and explore: clean and explore README.md
Clean and explore → Feature selection: prepare data README.md
Feature selection → Model selection: provide features README.md
Model selection → Model analysis: provide predictions src/analise_modelos/analise_modelos.py
Model analysis → Verification: evaluate results README.md
Verification → Model interpretation: validate first src/interpretacao_modelos/interpretacao_modelos.py
Model selection → Operational scenarios: inform scenarios
Feature selection → Model protocol: use rules src/feature_selection/feature_selection.py
Model selection → Model protocol: use protocol src/selecao_algoritmos/selecao_algoritmos.py
Model analysis → Model protocol: use metrics src/analise_modelos/analise_modelos.py
Verification → Model protocol: use protocol src/verificacao_validacao/verificacao_validacao.py
Model interpretation → Model protocol: load model rules src/interpretacao_modelos/interpretacao_modelos.py
Train model → Race files: read inputs
Train model → Model protocol: use model recipe
Race files → Batch prediction: provide races
Batch prediction → Race features: build features pipeline/prever.py
Race features → Recipe recalculation: calculate recipe pipeline/features_corrida.py
Batch prediction → Prediction files: write predictions
Operator → Race builder: assemble race README.md
Race builder → Simulation engine: validate and score app_streamlit/app.py
Race builder → Basket display: render display app_streamlit/app.py
Simulation engine → Simulator data: use material data
Simulation engine → Recipe recalculation: reuse recipe logic comum/receita.py
Simulation engine → Race builder: return risk app_streamlit/app.py

Drawn by GitDiagram on November 15, 2026 from the repository's file tree, README and source files. Read as Markdown ·