Predicting mixture concentration with pquad

This example shows how to plot predictions from the model

The parity plot for the mixtures where concentrations are known is shown in figure 1 and the plot of concentration with time for the experimental spectra from reacting systems are shown in figure 2 and 3 for different starting concentrations

from pquad import IR_Results
from pquad import get_defaults
from pquad.plotting_tools import set_figure_settings

Loading data

First, we’ll get the default data and load it into pquad. It automatically fits the model to the data in pure_data_path

frequency_range, pure_data_path, mixture_data_path, reaction_data_path = get_defaults()
deconv = IR_Results(4, frequency_range, pure_data_path)
deconv.set_mixture_data(mixture_data_path)
deconv_reaction = IR_Results(4, frequency_range, pure_data_path)
deconv_reaction.set_mixture_data(reaction_data_path, contains_concentrations=False)

Set figure settings

figure_folder='fit'
set_figure_settings('presentation')

Plot parity plot

Make parity plot with the 95% prediction intervals

deconv.plot_parity_plot(figure_folder)

Out:

R2 of mixed prediction: 0.9848041280249827
RMSE of mixed prediction: 0.019192974474399493
Max Error mixed prediction: 0.04810819524389309
C:\Users\lansf\Box Sync\Synced_Files\Coding\Python\Github\pQUAD\pquad\pquad.py:714: UserWarning: Matplotlib is currently using agg, which is a non-GUI backend, so cannot show the figure.
  plt.show()

Plot reaction plots

Make reaction plots with the 95% prediction intervals

deconv_reaction.get_reaction_figures(figure_folder)

• 
• 

Out:

C:\Users\lansf\Box Sync\Synced_Files\Coding\Python\Github\pQUAD\pquad\pquad.py:506: UserWarning: Matplotlib is currently using agg, which is a non-GUI backend, so cannot show the figure.
  plt.show()
C:\Users\lansf\Box Sync\Synced_Files\Coding\Python\Github\pQUAD\pquad\pquad.py:506: UserWarning: Matplotlib is currently using agg, which is a non-GUI backend, so cannot show the figure.
  plt.show()