READ ME File For Dataset for publication "Predictive Prosthetic Socket Design. Part 1: population-based evaluation of transtibial prosthetic sockets by FEA-driven Surrogate Modelling"

Dataset DOI: 10.5258/SOTON/D0484

ReadMe Author: J. W. Steer

This dataset supports the publication:
AUTHORS
TITLE
JOURNAL
PAPER DOI IF KNOWN


This contains instructions for opening the .pkl file containing the data behind the figures. The .pkl file is strutured using dictionaries

Opening the data file
---------------------

The data file is encoded using the pickle package in python 3.6. The data is all stored within numpy arrays. The authors recommend downloading the anaconda distribution of python. 

"""
import pickle
import numpy
with open('BMMMB_2019_data.pkl', 'rb') as f:
	Data = pickle.load(f)
"""

Figure 4
--------

Sensitivity analysis of the surrogate models relative to 75 observed data points from the FE simulations 

regions = ['Residuum tip pressure, kPa', 'Fibula head pressure, kPa',
		   'Tibial tuberosity pressure, kPa', 'Posterior calf shear, kPa', 'Distal tibia strain, %']

Within each region exists the data for each row in the figure. 'observed' refers to the 75 numbers from the FEA simulations, the numbers correspond to the number of training data points used to create the surrogate model
		   
data = ['input', 'observed', 25, 50, 100, 150, 200]	

Example
Data['Figure 4']['Residuum tip pressure, kPa']['observed']
Returns a numpy array of the observed data 


Figure 5
--------

Pressure, shear and soft tissue strain in key regions when varying residuum length and profile. 

First set of dictionary keys correspond to the region of the residuum 

regions = ['Residuum tip', 'Fibula head',
		   'Tibial tuberosity', 'Posterior calf', 'Distal tibia']

The data returned is a 50 x 50 x 3 array, corresponding to the X, Y and Z values in each plot 

'input' returns a list of the variables used to sample the surrogate model, these are either fixed numbers or variables

For figure 5, the input array [0.375, 0.375, 0.375, 'x', 'y', 0.5, 0.5]

Example
Data['Figure 5']['Residuum tip'][:, :, 2]
Returns a 50 x 50 array of predicted values at the residuum tip within the plot 

Figure 6
--------

Soft tissue strain for a 1% press fit socket when varying residuum length, residuum profile, soft tissue strain and tibia length. 

First set of dictionary keys correspond to the tissue modulus

modulus = [35, 40, 45, 50, 55]

Second set of dictionary keys correspond to the tibia length

tibia = [-15, 0, 15, 30]

The data returned is a 50 x 50 x 3 array, corresponding to the X, Y and Z values in each plot 

For figure 6, the input array is [0.375, 0.375, 0.375, 'x', 'y', 'mod', 'tib']

Example
Data['Figure 6'][35][0][:, :, 2] 
Returns a 50 x 50 array of the predicted values of soft tissue strain for a 35kPa soft tissue with 0% tibia length

Figure 7
--------

Pressure, shear and soft tissue strain in key regions when varying socket design for the 4 individuals 

First set of dictionary keys correspond to the individual 

Individual = [1, 2, 3, 4] 

Second set of dictionary keys correspond to the region 

regions = ['input', 'Residuum tip', 'Fibula head',
		   'Tibial tuberosity', 'Posterior calf', 'Distal tibia']

The data returned is a 50 x 50 x 3 array, corresponding to the X, Y and Z values in each plot 		   

For figure 7, the input array is ['x', ('x'+'y')/2, 'y', 'len', 'pro', 'mod', 'tib']

Example
Data['Figure 7'][1]['Residuum tip'][:, :, 2]
Returns a 50 x 50 array of the pressure at the residuum tip as the socket design varies 

Figure 8
--------

Pressure distribution 

Individual = [1, 2, 3, 4] 

Socket = ['input', -1, 1, 3, 5]

Returns a 1000 x 4 array corresponding to the node locations (first 3 columns) and corresponding pressure (final column) 

Example
Data['Figure 8'][1][3][:, 1] 
Returns a 1000 x 1 array of the pressure at each node for individual 1 with a 3% socket design


Information about geographic location of data collection: University of Southampton, U.K.

Licence: CC BY

Related projects:
ADD IN


Date that the file was created: 15 March 2019