MomentBalance is a professional research tool that creates statistically balanced experimental groups from CSV measurement data using high-performance parallel optimization algorithms. Core Functionality

CSV Data Import: Upload files with subject IDs and measurement values (volume in mm³, weights, biomarkers, etc.)
Dual Optimization: Balances both group means AND standard deviations simultaneously—crucial for reducing confounding variables
Parallel WASM Processing: WebAssembly-compiled Rust algorithm with Web Worker pool for true multi-core parallelism
Simulated Annealing Algorithm: Sophisticated probabilistic optimization with configurable quality levels (1-100, yielding 1-10 million iterations)
Flexible Group Configuration: Create 2-10 balanced groups with automatic size distribution
Real-time Data Preview: Instant validation and preview of uploaded CSV data

Advanced Features

High-Performance Computing: 20-25× faster than pure JavaScript through compiled WASM + parallel execution
Multi-Core Utilization: Automatically scales to available CPU cores for optimal performance
Multilingual Interface: Full support for Canadian English, Canadian French, German, Italian, and Japanese with automatic browser locale detection
Statistical Validation: Displays fitness scores, variance metrics, and balance quality assessments
Progress Tracking: Real-time progress feedback with completion notification
Data Persistence: Automatically saves settings between sessions via browser.storage.sync
Semantic UI: Clean, accessible interface with automatic dark/light theme using Pico CSS Classless framework
Export-Ready Results: Detailed group assignments with individual measurements and statistical summaries in CSV format

Research Applications

Animal Studies: Balance tumour volumes, body weights, biomarker levels
Clinical Trials: Create matched patient groups based on baseline measurements
Agricultural Research: Balance soil samples, plant measurements, treatment plots
Manufacturing QC: Balance product batches, material properties, testing samples
Educational Research: Create equivalent student groups for intervention studies

Technical Specifications

Architecture: Rust/WASM computation engine with JavaScript UI layer
Parallelization: Worker pool distributing independent optimization runs across CPU cores
Optimization Target: Minimizes variance of group averages + variance of group standard deviations
Algorithm: Greedy initialization followed by temperature-controlled simulated annealing
Complexity Scaling: Iterations scale logarithmically with problem size (log₁₀(specimens × groups))
Quality Levels: 10× ≈ 1M iterations, 100× ≈ 10M iterations with complexity adjustment
Package Size: ~0.11 MB total—includes WASM engine, UI framework, 5 language files, and all dependencies (~1/30th a typical photo)
Performance: 3-4× speedup from WASM compilation, 6-7× from parallelization (typical 8-core system)
Data Format: Simple CSV (comma-separated, no quotes required)
Localization: Locale-aware number formatting (decimal separators, thousands grouping)

Why MomentBalance? Traditional random assignment often creates unbalanced groups, introducing confounding variables that reduce statistical power. MomentBalance ensures your experimental groups start with equivalent statistical properties, improving the validity of your analyses and reducing the number of subjects needed to detect effects. The dual optimization of both means and standard deviations is particularly important when baseline variability differs between potential group assignments—a common issue in biological research where individual variation is high. The parallel WASM architecture enables processing larger datasets and higher quality settings than JavaScript-based tools, completing typical analyses in seconds rather than minutes. Perfect for researchers, statisticians, and quality control professionals requiring rigorous experimental design with reproducible group assignments.

Ответ разработчика