Simple Single Objective

In [1]:

import obsidian
print(f'obsidian version: ' + obsidian.__version__)

import pandas as pd
import plotly.express as px
import plotly.io as pio
pio.renderers.default = "plotly_mimetype+notebook"

obsidian version: 0.8.0

Set up parameter space and initialize a design¶

In [2]:

from obsidian import Campaign, Target, ParamSpace, BayesianOptimizer
from obsidian.parameters import Param_Categorical, Param_Ordinal, Param_Continuous

In [3]:

params = [
    Param_Continuous('Temperature', -10, 30),
    Param_Continuous('Concentration', 10, 150),
    Param_Continuous('Enzyme', 0.01, 0.30),
    Param_Categorical('Variant', ['MRK001', 'MRK002', 'MRK003']),
    Param_Ordinal('Stir Rate', ['Low', 'Medium', 'High']),
    ]

X_space = ParamSpace(params)
target = Target('Yield', aim='max')
campaign = Campaign(X_space, target, seed=0)
X0 = campaign.initialize(m_initial = 10, method = 'LHS')

X0

Out[3]:

	Temperature	Concentration	Enzyme	Variant	Stir Rate
0	8.0	17.0	0.1405	MRK001	Medium
1	12.0	143.0	0.1695	MRK003	Medium
2	4.0	101.0	0.2855	MRK002	High
3	28.0	87.0	0.1115	MRK002	Low
4	-4.0	115.0	0.2275	MRK001	Low
5	-8.0	73.0	0.0825	MRK002	Medium
6	20.0	129.0	0.0535	MRK001	High
7	24.0	31.0	0.2565	MRK002	Medium
8	16.0	59.0	0.1985	MRK003	High
9	0.0	45.0	0.0245	MRK003	Low

Collect results (e.g. from a simulation)¶

In [4]:

from obsidian.experiment import Simulator
from obsidian.experiment.benchmark import shifted_parab

simulator = Simulator(X_space, shifted_parab, name='Yield', eps=0.05)
y0 = simulator.simulate(X0)
Z0 = pd.concat([X0, y0], axis=1)

campaign.add_data(Z0)
campaign.data.sample(5)

Out[4]:

	Temperature	Concentration	Enzyme	Variant	Stir Rate	Yield	Iteration
Observation ID
5	-8.0	73.0	0.0825	MRK002	Medium	86.196333	0
2	4.0	101.0	0.2855	MRK002	High	46.270422	0
3	28.0	87.0	0.1115	MRK002	Low	60.288919	0
4	-4.0	115.0	0.2275	MRK001	Low	63.082417	0
1	12.0	143.0	0.1695	MRK003	Medium	44.280131	0

In [5]:

from obsidian.plotting import visualize_inputs

fig = visualize_inputs(campaign)

Fit the optimizer and visualize results¶

In [6]:

campaign.fit()

GP model has been fit                       to data with a train-score of: 1 for response: Yield

In [7]:

from obsidian.plotting import parity_plot, factor_plot, optim_progress

In [8]:

parity_plot(campaign.optimizer)

In [9]:

factor_plot(campaign.optimizer, feature_id=0)

Optimize new experiment suggestions¶

In [10]:

X_suggest, eval_suggest = campaign.suggest(m_batch=3)

In [11]:

df_suggest = pd.concat([X_suggest, eval_suggest], axis=1)
df_suggest

Out[11]:

	Temperature	Concentration	Enzyme	Variant	Stir Rate	Yield (pred)	Yield lb	Yield ub	f(Yield)	aq Value	aq Value (joint)	aq Method
0	-10.000000	10.0	0.131014	MRK001	Low	104.655575	97.694464	111.616687	2.349349	-0.380345	-0.292665	NEI
1	-10.000000	10.0	0.135985	MRK001	High	101.870533	94.441592	109.299474	2.195901	-0.612096	-0.292665	NEI
2	-1.227879	10.0	0.114474	MRK002	Low	101.243266	97.004850	105.481684	2.161341	-0.709565	-0.292665	NEI

Collect data at new suggestions¶

In [12]:

y_iter1 = pd.DataFrame(simulator.simulate(X_suggest), columns = ['Yield'])
Z_iter1 = pd.concat([X_suggest, y_iter1, eval_suggest], axis=1)
campaign.add_data(Z_iter1)
campaign.data.tail()

Out[12]:

	Temperature	Concentration	Enzyme	Variant	Stir Rate	Yield	Iteration	Yield (pred)	Yield lb	Yield ub	f(Yield)	aq Value	aq Value (joint)	aq Method	Yield (max) (iter)
Observation ID
8	16.000000	59.0	0.198500	MRK003	High	49.876574	0	NaN	NaN	NaN	NaN	NaN	NaN	NaN	92.339807
9	0.000000	45.0	0.024500	MRK003	Low	81.228810	0	NaN	NaN	NaN	NaN	NaN	NaN	NaN	92.339807
10	-10.000000	10.0	0.131014	MRK001	Low	92.196849	1	104.655575	97.694464	111.616687	2.349349	-0.380345	-0.292665	NEI	95.000467
11	-10.000000	10.0	0.135985	MRK001	High	74.499468	1	101.870533	94.441592	109.299474	2.195901	-0.612096	-0.292665	NEI	95.000467
12	-1.227879	10.0	0.114474	MRK002	Low	95.000467	1	101.243266	97.004850	105.481684	2.161341	-0.709565	-0.292665	NEI	95.000467

Repeat as desired¶

In [13]:

for iter in range(3):
    campaign.fit()
    X_suggest, eval_suggest = campaign.suggest(m_batch=3)
    y_iter = pd.DataFrame(simulator.simulate(X_suggest))
    Z_iter = pd.concat([X_suggest, y_iter, eval_suggest], axis=1)
    campaign.add_data(Z_iter)

GP model has been fit                       to data with a train-score of: 1 for response: Yield
GP model has been fit                       to data with a train-score of: 1 for response: Yield
GP model has been fit                       to data with a train-score of: 1 for response: Yield

In [14]:

optim_progress(campaign, color_feature_id = 'aq Value')

Analyze using Explainer¶

In [15]:

from obsidian.campaign import Explainer

In [16]:

exp = Explainer(campaign.optimizer)
exp.shap_explain(n=500)

  0%|          | 0/500 [00:00<?, ?it/s]

In [17]:

exp.shap_summary()

Out[17]:

In [18]:

exp.shap_pdp_ice(ind = 0, ice_color_var = 2)

Out[18]: