CNT-Polymer Nanocomposite Conductivity and Piezoresistivity Prediction Data and Models

Kavan Nailesh Shah's dataset contains source code, training data, and ML models for predicting multi-functional properties of CNT-polymer nanocomposites. It supplements a manuscript submitted to Computational Materials Science. The 126.4 MB repository includes files in PY, TXT, IPYNB, PTH, CSV, and RTF formats.

Use Cases

• Train neural networks to predict effective electrical conductivity based on statistical volume element (SVE) features.
• Predict piezoresistivity of nanocomposites using principal component scores from two-point correlation functions.
• Compare the predictive power of auto-correlations versus two-point cluster and blocking functions for material properties.
• Reproduce the study's results on structure-property linkages for CNT-polymer composites.

Strengths

• Includes full source code and trained ML models (PTH files) for reproducibility.
• Dataset size is 126.4 MB, suggesting substantial supporting material.
• License is GPL 3.0+, allowing for open modification and redistribution.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Description metadata is limited; actual data quality requires manual inspection after download.

Provenance

Source

figshare

Collection Method

Data generated from statistical volume elements (SVEs) of CNT-polymer nanocomposite microstructures.

Freshness

Last updated 2026-05-18 20:21:53; freshness should be verified.