Final-Year-Project

The algorithms for generation, visualisation, and AE-based classification of local compaction data are included as Jupyter notebooks.

Local compaction data generation (using window sizes in range 2-50)
Local compaction plot visualisation
a. Feature selection i - Random forest classifier training and analysis // b. Feature selection ii - Multilayer perceptron classifier training and weight extraction + visualisation // c. Feature selection iii - Estimation of Kolmogorov complexity using the gzip algorithm //
Autoencoder training and analysis - filtered data
Visualisation of data projection onto parametrised latent layer + clustering: using DBSCAN. Analysis of clusters obtained.

Also included are the local compaction plots for all window sizes (LCPs) and the Supplementary Figures SF1-SF3.

Name		Name	Last commit message	Last commit date
Latest commit History 14 Commits
01_Local compaction analysis - LC data generation.ipynb		01_Local compaction analysis - LC data generation.ipynb
02_LCPs.ipynb		02_LCPs.ipynb
03a_Feature selection - RF feature importance analysis.ipynb		03a_Feature selection - RF feature importance analysis.ipynb
03b_Feature selection - MLP classifier weights.ipynb		03b_Feature selection - MLP classifier weights.ipynb
03c_Feature selection - Kolmogorov complexity.ipynb		03c_Feature selection - Kolmogorov complexity.ipynb
04a_Autoencoder_training_fil.ipynb		04a_Autoencoder_training_fil.ipynb
05_Latent layer visualisation and clustering.ipynb		05_Latent layer visualisation and clustering.ipynb
LCPs_all.png		LCPs_all.png
README.md		README.md
Supplementary Figures.pdf		Supplementary Figures.pdf
myc_091-160_D132-H_stripped.prmtop		myc_091-160_D132-H_stripped.prmtop
myc_091-160_stripped.prmtop		myc_091-160_stripped.prmtop

Provide feedback