AutoML for Classification Model Screening

The following are the steps for using AutoML for a classification task:

Provide an input pandas dataframe consisting of two columns:
- SMILES strings
- target class values
If featurization = True, Molecules are represented as:
- Coulomb matrix
- RDKit Morgan (ECFP6 1024-bit) fingerprints
- MACCs keys
- RDKit hashed topological torsion
- RDKit molecular descriptors (all)
- Mordred molecular descriptors
If featurization=False, a feature set must be provided by the user.
Screens through various classifier models:
- Single-core (runs regardless of multi_core value):
- Multi-core(set multi_core=True):
  - RandomForestClassifier
  - ChemML_MLP (if n_datapoints>500)
  - (optional) XGBClassifier
  - (optional) LightGBM

Yields ‘n-best’ models, with optimized hyperparamters.

The output_dir argument specifies a folder that will be created wherever the code is run from, containing the following directory tree:

 <output_dir>/
 |-- output.log
 |-- scores.csv
 |-- best_model/
 |   |-- model.pkl # most non-MLP models
 |   |-- model_chemml_model.json # MLP case (plus possible extra model files)
 |   |-- scaler.pkl # present when scaler exists for chosen feature_key
 |   |-- feature_order.json
 |   |-- metadata.json
 |-- .model_screener_artifacts/
 |   |-- <run_key_1>/
 |        |-- model.pkl # or MLP artifact files
 |   |-- <run_key_2>/
...

Load your data

We are using ChemML’s in-built dataset to demonstrate the minimum requirements to run AutoML.

[1]:

import pandas as pd
from chemml.datasets import load_organic_density

[2]:

molecules, target, dragon_subset = load_organic_density()
df=pd.concat([molecules, target], axis=1)
df = df.sample(200, random_state=42)
df

[2]:

	smiles	density_Kg/m3
361	c1scc(n1)c1ncc(s1)c1c(ccc2c1cccc2)c1cncs1	1328.35
73	S1CC(SC(C1)C1CCCC1)c1cocc1c1ccccc1	1143.80
374	N1CNC(NC1)c1sccc1C1(CSCCS1)c1cccs1	1351.33
155	c1nc(c(s1)c1nsnc1)c1csc(n1)c1cccs1	1466.57
104	Oc1ccc2c(c1c1cocc1)c(ccc2)c1ccccn1	1207.67
...	...	...
248	Oc1ncc(c(c1)C)c1cnccn1	1195.16
145	Sc1cc2ccccc2c(c1)c1cnccc1c1cnccn1	1213.41
92	c1sc(c(c1)c1ncccc1c1nsnc1)c1ccc2c(c1)cccc2	1264.16
152	N1CNC(NC1)(N1CNC(NC1)c1cncs1)c1cccc2c1cccc2	1270.40
367	C1CSC(CS1)C1(CSC(CS1)c1ccc2c(c1)cccc2)c1cnccn1	1268.28

200 rows × 2 columns

[3]:

# Labeling data for classification instead of regression
# Note: This is just for demonstration purposes. In practice, you would want to use a regression model for predicting density.
df['density_class'] = pd.qcut(df['density_Kg/m3'], q=3, labels=range(3))

Run autoML for a classification task

[4]:

from chemml.autoML import ModelScreener

output_dir = "automl_results_class"

MS = ModelScreener(df, target="density_class", featurization=True, cache_features=False, smiles="smiles",
                   screener_type="classifier", output_dir=output_dir)
scores = MS.screen_models(n_best=10, multi_core=False)



-------------------------ChemML AutoML Model Screening-------------------------

-------------------------Model screening started at Mon Aug 31 01:28:19 2026-------------------------



Converting SMILES to ChemML Molecule objects...

Converting SMILES to ChemML Molecule objects: 100%|██████████| 200/200 [00:06<00:00, 28.92it/s]


Generating Coulomb matrix representation...

featurizing molecules in batches of 12 ...
200/200 ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 14s 69ms/step
Merging batch features ...    [DONE]

Generating RDKit fingerprint representations...

Generating Morgan Fingerprints: 100%|██████████| 200/200 [00:00<00:00, 10268.08molecule/s]
Generating MACCS Fingerprints: 100%|██████████| 200/200 [00:00<00:00, 998.62molecule/s]
Generating Hashed Topological Torsion Fingerprints: 100%|██████████| 200/200 [00:00<00:00, 8634.43molecule/s]
Generating Atom Pair Fingerprints: 100%|██████████| 200/200 [00:00<00:00, 9262.83molecule/s]


Generating RDKit descriptor representations...

Step 1: Normalizing molecule input

Step 2: Calculating RDKit descriptors: 100%|██████████| 200/200 [00:02<00:00, 95.34it/s]


Generating Mordred descriptor representations...

Step 1: Normalizing molecule input

Normalizing Molecule input: 100%|██████████| 200/200 [00:00<?, ?it/s]

Step 2/3: Optimizing geometry for 3D descriptor calculation

3D geometry already exists, loading geometry: 100%|██████████| 200/200 [00:00<?, ?it/s]

Step 3/3: Calculating Mordred descriptors


100%|██████████| 200/200 [00:06<00:00, 31.55it/s]


Cleaning feature sets to remove constant, highly correlated, low-variance, and complex features...

Feature set 'CoulombMatrix' cleaned: 25 features retained.

Feature set 'morganfingerprints_radius3' cleaned: 252 features retained.

Feature set 'MACCS_radius3' cleaned: 48 features retained.

Feature set 'hashedtopologicaltorsion_radius3' cleaned: 76 features retained.

Feature set 'hashedatompair_radius3' cleaned: 421 features retained.

Feature set 'rdkit_descriptors' cleaned: 90 features retained.

Feature set 'mord_descriptors' cleaned: 530 features retained.


-------------------------Run parameters-------------------------
  Featurization: True
  Multi_core: False
  Screener_type: classifier
  Number of datapoints: 200
  Note: Dataset <= 500 samples; MLP is excluded from screening due to inefficiency.

split done!
Single-core complete


------------------------- Screening complete for feature set CoulombMatrix, time taken: 16.359 seconds -------------------------

split done!
Single-core complete


------------------------- Screening complete for feature set morganfingerprints_radius3, time taken: 13.086 seconds -------------------------

split done!
Single-core complete


------------------------- Screening complete for feature set MACCS_radius3, time taken: 12.781 seconds -------------------------

split done!
Single-core complete


------------------------- Screening complete for feature set hashedtopologicaltorsion_radius3, time taken: 12.867 seconds -------------------------

split done!
Single-core complete


------------------------- Screening complete for feature set hashedatompair_radius3, time taken: 14.05 seconds -------------------------

split done!
Single-core complete


------------------------- Screening complete for feature set rdkit_descriptors, time taken: 12.805 seconds -------------------------

split done!
Single-core complete


------------------------- Screening complete for feature set mord_descriptors, time taken: 16.855 seconds -------------------------

[5]:

scores

[5]:

	Model	Accuracy	Recall	Precision	F1-score	time(seconds)	parameters	Feature	run_key
27	LogisticRegression	0.8	0.793651	0.844444	0.797587	9.955029	[{'C': 0.1, 'class_weight': None, 'dual': Fals...	mord_descriptors	LogisticRegression_mord_descriptors_324dd7b551...
26	KNeighborsClassifier	0.75	0.746032	0.811111	0.758371	6.259871	[{'algorithm': 'auto', 'leaf_size': 30, 'metri...	mord_descriptors	KNeighborsClassifier_mord_descriptors_1e7fa7c1...
24	DecisionTreeClassifier	0.7	0.690476	0.763889	0.694444	3.405225	[{'ccp_alpha': 0.0, 'class_weight': None, 'cri...	mord_descriptors	DecisionTreeClassifier_mord_descriptors_6a4d98...
19	LogisticRegression	0.65	0.642857	0.662963	0.646562	7.150034	[{'C': 42.16315789473685, 'class_weight': None...	hashedatompair_radius3	LogisticRegression_hashedatompair_radius3_c444...
17	SVC	0.6	0.595238	0.604762	0.597691	2.851144	[{'C': 0.0037926901907322525, 'break_ties': Fa...	hashedatompair_radius3	SVC_hashedatompair_radius3_ef5a1fb731b9493682d...
15	KNeighborsClassifier	0.55	0.563492	0.541667	0.545788	5.889152	[{'algorithm': 'auto', 'leaf_size': 30, 'metri...	hashedtopologicaltorsion_radius3	KNeighborsClassifier_hashedtopologicaltorsion_...
16	DecisionTreeClassifier	0.5	0.507937	0.541667	0.505628	1.434301	[{'ccp_alpha': 0.0, 'class_weight': None, 'cri...	hashedatompair_radius3	DecisionTreeClassifier_hashedatompair_radius3_...

Validating AutoML-generated model

Since ChemML v1.3.4, AutoML outputs the best model and required scaler and feature order as easy-to-import files. A quick demonstration of the model loading workflow is given below:

[6]:

import pickle
import json

with open(f'{output_dir}/best_model/metadata.json', 'r') as f:
    metadata = json.load(f)
with open(f'{output_dir}/best_model/explanation.json', 'r') as f:
    explanation = json.load(f)
with open(f'{output_dir}/best_model/feature_order.json', 'r') as f:
    feature_order = json.load(f)
with open(f"{output_dir}/best_model/model.pkl", "rb") as f:
    loaded_model = pickle.load(f)
with open(f"{output_dir}/best_model/scaler.pkl", "rb") as f:
    loaded_scaler = pickle.load(f)

Metadata isn’t necessary for model validation, but gives us a lot of context from the AutoML screening workflow.

[7]:

metadata

[7]:

{'artifact_schema_version': 1,
 'model_name': 'LogisticRegression',
 'feature_key': 'mord_descriptors',
 'run_key': 'LogisticRegression_mord_descriptors_324dd7b551dd4d0e89563c3fa79868b8',
 'screener_type': 'classifier',
 'parameters': {'C': 0.1,
  'class_weight': None,
  'dual': False,
  'fit_intercept': True,
  'intercept_scaling': 1,
  'l1_ratio': 0.0,
  'max_iter': 100,
  'n_jobs': None,
  'penalty': 'deprecated',
  'random_state': None,
  'solver': 'sag',
  'tol': 0.0001,
  'verbose': 0,
  'warm_start': False},
 'n_folds': 5,
 'n_datapoints': 200,
 'metrics': {'Accuracy': 0.8,
  'Recall': 0.7936507936507936,
  'Precision': 0.8444444444444444,
  'F1-score': 0.7975867269984916,
  'time(seconds)': 9.955028772354126},
 'model_serializer': 'pickle',
 'model_artifact_dir': 'model_artifact',
 'scaler_file': 'scaler.pkl',
 'feature_order_file': 'feature_order.json',
 'timestamp': 'Mon Aug 31 01:30:34 2026'}

Explanation gives the user a list of feature importances alongside feature names; while not necessary for validation, this is useful to understand what features the model is actually using.

[ ]:

explanation

{'model_type': 'LogisticRegression',
 'explanation_type': 'linear.coef_',
 'importance_values': {'BCUTs-1l': 0.12080837896578821,
  'ATSC6s': 0.11705360020911237,
  'BCUTdv-1l': 0.11643211939895474,
  'MATS6s': 0.11176801885427924,
  'MIC0': -0.10732155496149466,
  'BCUTZ-1h': -0.10718923909200691,
  'VSA_EState7': 0.10557335217812619,
  'SM1_DzZ': -0.10481593783264181,
  'MIC2': -0.10150151886248547,
  'MDEC-22': 0.10067185163430138,
  'PEOE_VSA4': -0.10042496013838315,
  'Mor13': 0.09870373105961783,
  'IC0': -0.09707281423378299,
  'C2SP3': 0.09600380966243331,
  'Mor19': 0.09434981175135433,
  'IC2': -0.09216564338412887,
  'Mor23v': -0.08944319176524317,
  'WNSA4': 0.08883671204663847,
  'Mor11m': 0.08859302230665116,
  'nHetero': -0.08830107172063842,
  'Mor21': -0.0877983270264638,
  'nS': -0.0852109498551405,
  'MATS1s': 0.0849868365554287,
  'ATSC8i': -0.08486929913376605,
  'Mor11v': 0.0843377479785362,
  'PEOE_VSA6': 0.08371084437200652,
  'GATS3Z': 0.08314941767103975,
  'ZMIC4': -0.08291889991449908,
  'SlogP_VSA11': -0.08264076194322785,
  'Mor22': 0.0811945672039007,
  'AATSC0Z': -0.08025587707186668,
  'AATSC2Z': 0.07969179763267753,
  'ATSC0Z': -0.07938387796472843,
  'GATS1s': -0.0778363152307057,
  'PNSA4': 0.07711466587068334,
  'WNSA3': 0.07702985845556001,
  'SMR_VSA5': 0.07635130453924953,
  'ATS0Z': -0.07621573428909163,
  'MIC1': -0.07537027853437128,
  'Mor07v': 0.07493426507957587,
  'ATSC2Z': 0.07423512895464822,
  'Mor17': -0.07265706116251208,
  'SM1_Dzse': -0.07258756393687768,
  'Mor23m': -0.07243901248933617,
  'AMID_h': -0.0715074895084047,
  'Mor18': -0.07134081353269153,
  'FNSA2': 0.07098238606356791,
  'DPSA3': -0.07086776819047452,
  'WNSA5': 0.07072208522949826,
  'NaaS': -0.07011039127881565,
  'Mor21m': -0.06984160210315588,
  'Mor23': -0.06979633396704292,
  'AATS5Z': -0.06776667525244524,
  'DPSA4': -0.06743265563863299,
  'AATS6Z': -0.0671288925307278,
  'ATSC4c': -0.06709979041594832,
  'Mor31': 0.06705804301418883,
  'DPSA1': 0.06667074035047829,
  'PNSA2': 0.06665612606226767,
  'PNSA3': 0.06598975001498042,
  'Mor07': 0.06558397866509244,
  'ATSC6v': -0.06527858478714987,
  'AATS0Z': -0.06513692171231619,
  'RNCS': 0.06500833397631156,
  'VSA_EState6': 0.06467998846396998,
  'ATSC1s': 0.0639567175973236,
  'SRW05': -0.06385335676797899,
  'Mor03m': 0.0635536906752689,
  'ZMIC2': 0.0633770766472029,
  'ATSC1dv': 0.06315167855356948,
  'Mor07m': 0.06299027299502238,
  'PNSA1': -0.06284026238877223,
  'MATS1c': 0.0627341979463412,
  'SlogP_VSA5': 0.0625480201773812,
  'GATS3p': 0.062294692738831585,
  'Mor27': -0.06183006226945349,
  'Mor06p': -0.061320930862255645,
  'NsCH3': 0.061107468780475446,
  'AATSC7i': 0.0609338990453993,
  'Mor25': 0.06088850374545029,
  'Mor16': 0.06065044433704577,
  'n5HRing': -0.06056563915935044,
  'ATSC6p': -0.060454239126411874,
  'PEOE_VSA9': -0.0600792995531491,
  'PNSA5': 0.05925226467000524,
  'ATSC1c': 0.05879170065770135,
  'EState_VSA4': -0.05878302306780468,
  'Xch-7dv': -0.058671009652759135,
  'Mor12': -0.05831698042454397,
  'Mor27v': -0.05783720073530536,
  'n6aRing': 0.05779768084307773,
  'Mor05': -0.056562456827765124,
  'AATS5p': -0.056103765160925564,
  'AATS6p': -0.05584058998601546,
  'nC': 0.055696487376654696,
  'nHRing': -0.05569562438544719,
  'NsSH': -0.05557082227059615,
  'Mor22m': 0.05546542693895479,
  'ATSC3c': 0.05545302844910022,
  'Mor24m': -0.05518521095380485,
  'GATS3s': 0.05507287368131927,
  'ATSC3s': -0.054405275267100005,
  'ATSC1pe': 0.053341518437240255,
  'Mor20': 0.05289003546006708,
  'AATS6se': -0.05263014767903107,
  'ATSC6Z': -0.05225501029632406,
  'Mor22v': 0.052208026618504474,
  'AATS1se': -0.052016972573772045,
  'WNSA1': -0.05171974088332188,
  'Mor27m': -0.05160604383234613,
  'Mor02m': -0.05150896285084032,
  'ATSC8p': -0.05140676549186002,
  'Mor08': 0.05065750034077404,
  'ATSC1d': 0.05049038574033142,
  'AATS7se': -0.050397534695051004,
  'GATS7dv': 0.050337125780684985,
  'AATS0p': -0.050133006029732294,
  'nH': 0.04994985338326599,
  'ATSC7pe': -0.0499325873419884,
  'ATSC7i': 0.049757529903058186,
  'Mor08m': -0.049655392893681344,
  'n6aHRing': 0.049326708043321135,
  'SMR_VSA6': -0.04875171896956202,
  'ATSC0c': -0.04853618150329924,
  'Mor06m': -0.04843296330364251,
  'MATS7se': -0.04825440075448658,
  'AATS0se': -0.04803193343483492,
  'NaaCH': 0.047451231845715536,
  'ATSC6d': -0.047438229513838054,
  'Xch-7d': -0.04672316245021522,
  'n5ARing': 0.04656107401997829,
  'GATS8c': 0.046445633614312684,
  'Mor03v': 0.04613276455370276,
  'ATSC2s': -0.045158917984958566,
  'AATSC6dv': 0.045088567124141644,
  'GATS2i': 0.0449961983869565,
  'SssssC': -0.04484900692779872,
  'Xch-6dv': -0.044822868571114144,
  'Mor19v': 0.044568215687565874,
  'PPSA5': -0.044566700343472855,
  'SlogP_VSA2': -0.04385382135295167,
  'Mor13m': -0.043827443270466844,
  'PPSA1': 0.04372327809832466,
  'GATS5se': 0.04369325110545239,
  'AATS6v': -0.04343297632486252,
  'AATSC0i': -0.04341791286884424,
  'GATS7i': -0.0434061357080254,
  'EState_VSA5': 0.043374799734754096,
  'AATS6d': -0.0433590548614338,
  'VSA_EState4': -0.04324451352340742,
  'GATS5s': 0.043077255568700526,
  'IC3': -0.04304824612498354,
  'GATS1c': -0.04268478018796047,
  'BCUTi-1h': -0.042563074562621574,
  'GATS8se': 0.04234499544350699,
  'Mor04': 0.042174025066053104,
  'ATSC1i': 0.042123078096283836,
  'Mor11': 0.04183428134082267,
  'Mor09v': -0.04139005913449453,
  'PBF': -0.041357787060382276,
  'ATS1s': 0.041118530026729085,
  'WPSA4': -0.04103054559761104,
  'ATSC1v': 0.04097175567563154,
  'Mor06v': -0.04088194743414806,
  'PPSA3': -0.040781104102206964,
  'TSRW10': -0.04067851727584217,
  'nAtom': 0.04061022541330121,
  'ATSC5d': -0.04057738727442518,
  'AATSC1dv': 0.040422250329272824,
  'PPSA4': -0.040403788344096726,
  'GATS8i': 0.04030302886012403,
  'ATSC8v': -0.03998086414854392,
  'FilterItLogS': -0.039921638053251565,
  'ATS1Z': -0.03971944245657374,
  'GATS8d': 0.039695789372516,
  'Mor09m': -0.03913069799730909,
  'EState_VSA7': -0.03912183849300865,
  'BCUTdv-1h': 0.03900251000427328,
  'MATS8se': -0.03868718084008164,
  'PEOE_VSA3': -0.038544662912747045,
  'EState_VSA8': 0.03832646356895805,
  'GGI5': -0.03818453197788489,
  'VSA_EState1': -0.03813749920248062,
  'AATS4Z': -0.03793591279064621,
  'C2SP2': 0.037918849523622865,
  'Mor28v': 0.03791048872968623,
  'ATSC2p': 0.0377302251776739,
  'EState_VSA9': -0.03772536321427576,
  'AATSC8i': -0.03756741721389703,
  'ATSC7s': -0.03752247664142901,
  'GATS3pe': 0.03748397243724362,
  'MATS7pe': -0.037454316039981214,
  'Mor03p': 0.03742225801021257,
  'TIC0': -0.037091138743572956,
  'MOMI-Z': -0.03673544730428269,
  'ATSC2c': -0.03659842931031411,
  'WPSA3': -0.03655098660898325,
  'AATS8Z': -0.036426949000269715,
  'GATS4c': 0.03640763506047627,
  'C3SP2': 0.03622097013264929,
  'n6Ring': 0.03600524256872238,
  'VSA_EState3': -0.03598617330057456,
  'AATS8se': -0.03598336238532537,
  'AATS8v': -0.03588540437424892,
  'Mor25v': 0.03575406686616382,
  'GATS6c': -0.03571323705637587,
  'AATS7dv': -0.035709233992754275,
  'SlogP_VSA4': 0.035595726333001766,
  'ATSC2i': -0.035533947698472214,
  'GATS7pe': 0.03544377091744612,
  'IC1': -0.035318856821838794,
  'GATS2Z': -0.03523082334691317,
  'AATSC5Z': -0.03518587363716325,
  'GATS8v': 0.03515408663806315,
  'GATS5c': 0.03499472480062851,
  'nN': -0.034973671072659474,
  'AATS3se': -0.03479487354198179,
  'Mor28': 0.03465975459211602,
  'ATSC1se': 0.03459650262766007,
  'Mor20v': 0.03455455953869654,
  'GATS3se': 0.03417833564230912,
  'AATS7d': -0.03379234214974467,
  'ATSC6se': 0.03368938958085084,
  'ATSC4s': 0.03348157704550467,
  'AETA_eta_F': 0.03344524366162438,
  'GGI3': 0.03333809347740654,
  'GATS7s': 0.03329907226614312,
  'SLogP': 0.033296158740393066,
  'ATSC5Z': -0.03291013612885785,
  'Mor14': 0.032494717094798266,
  'Mor02v': -0.03243454950353637,
  'ATSC1p': 0.03235723703221591,
  'GATS4d': 0.03213071756937961,
  'WPSA1': 0.03204081935416987,
  'ZMIC1': 0.03175504914419546,
  'GATS6s': -0.03174403296326133,
  'AATS7v': -0.03158880137632405,
  'GATS8dv': 0.03149243806228315,
  'BCUTv-1l': 0.03141600025632607,
  'Mor02': -0.03141313893835163,
  'Mor17m': -0.03108572632809317,
  'Mor31v': 0.031030703631038617,
  'AATS0v': -0.03094044861566445,
  'AATS8dv': -0.03089234179186471,
  'ATSC2d': 0.030859505152072943,
  'ETA_beta_ns_d': -0.030760271583760854,
  'AATS5se': -0.030673974188087906,
  'AATSC7Z': 0.03062008976467808,
  'GATS5dv': 0.030363024750366778,
  'BCUTd-1l': 0.03034751402930108,
  'AATS5v': -0.030054379260253988,
  'Mor30': 0.030026921537975884,
  'Mor12v': -0.02995494123464769,
  'C1SP3': -0.029850079987276748,
  'AATS0d': -0.029786400838106974,
  'ATSC6c': 0.029634235738143305,
  'AATS7p': -0.029579194629221896,
  'ATSC8dv': 0.02956838830959986,
  'GATS7se': 0.029560655514654085,
  'AATS7Z': -0.02938696938277561,
  'ATSC7p': 0.029294574241194842,
  'ATSC7dv': -0.02927189635074724,
  'Mor15m': 0.02924786240531492,
  'Xch-5d': -0.0292284675316169,
  'Mor26': 0.029227336506969362,
  'ATSC3se': 0.029090242864351516,
  'MATS8c': -0.028737757192897992,
  'GATS7c': 0.02871728591065137,
  'ATSC4Z': 0.02861463295868068,
  'GATS4dv': -0.028602795346020473,
  'SMR_VSA9': -0.028436211874360486,
  'ATSC0pe': -0.028174027562501442,
  'GATS8s': 0.028027497144830545,
  'Mor32': -0.027993098403816467,
  'ATSC5pe': -0.027865227096743512,
  'EState_VSA10': -0.027765215131674813,
  'Mor16m': -0.02747678915681787,
  'SpDiam_A': 0.02707751469458059,
  'GATS4Z': -0.027076555261340566,
  'NaaaC': 0.027051961639268735,
  'AETA_eta': -0.02635034188664906,
  'SlogP_VSA3': 0.026101049264122173,
  'GATS1v': 0.025984281294022943,
  'PEOE_VSA1': -0.02595518355265528,
  'GATS1i': -0.02579969730869756,
  'ATSC7se': -0.025749022568350342,
  'GATS6se': -0.024877375469614656,
  'GATS1dv': -0.02487193856883485,
  'ZMIC3': -0.02481406451167846,
  'ATSC0s': -0.024664162611399188,
  'Mor29m': 0.024635822452118733,
  'naHRing': -0.0244160224209537,
  'Mor26m': 0.024393142174503377,
  'PEOE_VSA11': -0.024158926935076683,
  'SsssCH': 0.024151614605519355,
  'AATS5d': -0.0240537394560784,
  'C3SP3': -0.023930094098212035,
  'ATS8Z': -0.023845978695860522,
  'Mor18m': 0.02372629202014074,
  'Mor14m': 0.023681010725991394,
  'ATSC4pe': 0.023497304830171806,
  'PEOE_VSA7': 0.02335863107276624,
  'ATSC3v': 0.023224391427887205,
  'Mor14v': 0.022849585533492225,
  'AATSC6i': 0.022733282370838992,
  'Mor10m': 0.022693229431218653,
  'PEOE_VSA10': 0.022612955260741616,
  'Mor03': 0.022574412132690302,
  'GATS7p': -0.022562967133813918,
  'AATS8d': -0.022506183812222098,
  'ATSC3d': 0.022434065641580275,
  'MATS2c': -0.022235910306878787,
  'Kier3': 0.02209691431108705,
  'Mor18v': -0.022034478423942447,
  'Mor20m': 0.022030408353566844,
  'AATSC1Z': -0.021798724701541445,
  'AATS6i': 0.021795803436324458,
  'ATSC5s': 0.021724251870350087,
  'GATS8p': 0.021561892868083674,
  'AATSC3Z': -0.02138593757099824,
  'AATS4v': -0.021380028623155453,
  'Mor30v': 0.02128891759160878,
  'nBondsS': 0.021245438965662014,
  'ATSC5v': 0.021165322577128286,
  'AATS6dv': -0.020618411361216192,
  'ATSC3Z': 0.02055258260850909,
  'GeomRadius': -0.020365194458673772,
  'GGI6': 0.020245840183482973,
  'GATS6i': -0.01996577030664135,
  'GATS4s': -0.019901593682738417,
  'Mor19m': 0.019894103328682643,
  'ATSC8Z': -0.0197225636172513,
  'nBase': -0.019564851553910717,
  'AATSC4Z': 0.01954478320106971,
  'Mor26v': 0.01944767438693513,
  'ATS0s': 0.019420890511106893,
  'Mor05m': 0.0191717524435837,
  'nO': -0.019102406157183102,
  'NssS': -0.019022275818320313,
  'Mor24v': -0.01899422049238081,
  'Mor15': -0.018963789164845295,
  'PEOE_VSA2': -0.01888627921356993,
  'AATS5i': 0.018883019070424507,
  'NaasC': -0.018860335924556344,
  'ATSC4se': -0.018718299423674763,
  'MATS8pe': -0.018713655414184422,
  'GATS2s': 0.01845801972633954,
  'SlogP_VSA10': -0.018453408599639877,
  'Mor15v': 0.018370474208481303,
  'AATSC3dv': -0.01816487281000892,
  'Mor16v': -0.018039991701816402,
  'ATS8dv': -0.01803103830420602,
  'bpol': -0.01790988535814143,
  'SpMAD_Dzv': -0.017662584935053235,
  'AATSC4dv': 0.017655581863441818,
  'ATSC0dv': 0.01759992703689625,
  'GeomDiameter': -0.017592270004226924,
  'VSA_EState5': -0.017575783706958507,
  'Mor24': -0.017247652356007122,
  'AATSC0dv': -0.01679405907743964,
  'GATS3i': 0.016729722877538884,
  'GATS3d': 0.016716326167183247,
  'nAromAtom': 0.01663544423871456,
  'ATSC8pe': 0.0165489757976493,
  'BCUTs-1h': -0.016375298303912638,
  'EState_VSA3': -0.01637158872863449,
  'Mor30m': 0.0162990875191024,
  'ATSC4dv': 0.016052266391098753,
  'AATS0dv': -0.01597193168906244,
  'ATSC5i': 0.01595376511830446,
  'Xc-5dv': -0.0159334533542992,
  'BCUTse-1h': 0.01581502820625964,
  'SM1_Dzv': 0.01567428400897486,
  'ATSC4i': -0.01555376531026534,
  'AATS8s': -0.015507668494879032,
  'ATS5Z': -0.015505966323840633,
  'AATS7i': 0.01529088377921347,
  'AATSC5i': -0.015206230067802503,
  'ABC': 0.01504900999678791,
  'ATSC1Z': -0.01494853102953502,
  'Mor17v': -0.014929294487106051,
  'Mor09': -0.0148456512922033,
  'Mor10': 0.014818335522133267,
  'ATSC4d': -0.014671718922032085,
  'EState_VSA6': -0.014620905427907517,
  'NaaN': -0.01461249488018046,
  'NssCH2': 0.014433326572594258,
  'AATSC5dv': 0.014361123743243328,
  'ATS4Z': -0.014295005564255907,
  'Mor04v': -0.014274482358727676,
  'Mor29v': 0.013942273803252875,
  'AATSC7v': 0.013927729782467571,
  'AATSC6v': -0.013845625977691081,
  'IC4': 0.013841369031042643,
  'ATSC6i': -0.013809963550853866,
  'GATS6d': -0.013278827188968063,
  'AATS5s': 0.01327335136899033,
  'AATSC4v': -0.013265763055328393,
  'AATSC4i': -0.013208798229842698,
  'AATS6s': 0.013111145876303718,
  'BCUTd-1h': 0.013039791446187431,
  'AATS1s': 0.01297712631644322,
  'MATS7s': -0.012819518397026159,
  'ATSC6dv': -0.012695458971319025,
  'ATSC2pe': 0.012627986175127386,
  'MATS3c': 0.012532211191348218,
  'GATS1se': -0.01226020248651081,
  'GATS4se': 0.012044764182006677,
  'ATS0p': -0.01202874740132981,
  'MOMI-X': -0.01195705188224151,
  'AATSC8dv': 0.011593901857097626,
  'EState_VSA2': 0.011485509005819157,
  'Mor29': 0.011442491495566209,
  'ATSC3pe': 0.011252494641212038,
  'GATS6Z': 0.011004069202544696,
  'MATS6Z': -0.010922885759840106,
  'SpMAD_Dzse': 0.01091386629500248,
  'AATSC3i': -0.010792789862865448,
  'VSA_EState2': -0.010669171906056407,
  'PPSA2': -0.010552607701175708,
  'NsssN': -0.01048360294504262,
  'NsOH': -0.010399404378233161,
  'GGI4': 0.010371577152687722,
  'ATSC8s': -0.010289387062335077,
  'HybRatio': -0.010203032282095544,
  'VSA_EState8': -0.01012299968417748,
  'Xc-3d': 0.01004582444826636,
  'NsssCH': 0.00987138365890545,
  'ATSC3i': 0.009860001349554158,
  'AATSC8v': -0.009851679397161897,
  'GATS5Z': 0.0098296033697443,
  'Mor10v': 0.009524195326703427,
  'RPCS': 0.009479618511522283,
  'GATS4i': 0.009434124915095588,
  'AATSC3v': -0.009250485551279448,
  'ATSC2dv': -0.009248887894812406,
  'n6HRing': 0.009220218970159056,
  'GATS1p': 0.00913905968461284,
  'ABCGG': 0.008846590790875604,
  'ATSC7Z': 0.008836035647976647,
  'ATSC5p': -0.008771074035330353,
  'ATSC6pe': 0.008741724958691814,
  'GATS7d': 0.008658437699936567,
  'Mor31m': -0.008631865746234656,
  'PEOE_VSA5': 0.008608479318108455,
  'NaaO': 0.008470508231880062,
  'Mor06': -0.008324047346850843,
  'AATSC5v': 0.008244623690123996,
  'Mor25m': 0.008210285307578077,
  'Mor08v': -0.008176492232599985,
  'TIC1': 0.008111812714411117,
  'TASA': -0.008076971938866709,
  'PEOE_VSA8': 0.007991740676410645,
  'AATSC0s': -0.007880053497496244,
  'AATS4d': -0.007847738060549957,
  'ATS3Z': -0.007752947283317865,
  'GATS5p': -0.007731696636388393,
  'ATSC3dv': 0.007701331853378726,
  'AATSC6Z': 0.007673583603162428,
  'GATS7Z': -0.007672085357791472,
  'GATS6dv': 0.007596712280292142,
  'GATS4p': 0.007556136439638323,
  'GeomShapeIndex': 0.007549975306883341,
  'GATS2p': -0.007521998694664529,
  'ATSC8d': 0.007378353981595479,
  'MATS8s': -0.007371633631964059,
  'GATS5d': 0.007259039134314974,
  'GATS3c': 0.006907057823903871,
  'AATS3i': 0.006784190636373167,
  'ATSC7v': -0.006647635199080438,
  'ATSC4v': -0.006415697640047022,
  'GATS6p': 0.006413182801058924,
  'TPSA': 0.006306411336226291,
  'GATS2d': 0.006004335795723908,
  'Diameter': 0.005906372134926628,
  'GATS2dv': 0.005632581173283952,
  'Mor13v': -0.005505364083768485,
  'Mor32v': -0.005309891043961399,
  'AATS4s': -0.004972038769412041,
  'ATSC8se': 0.00496770140295598,
  'SaasC': 0.0048910366323222255,
  'GATS2se': 0.004797693231150838,
  'GATS2c': -0.004746725781479923,
  'nARing': -0.004642425776856336,
  'SpAbs_Dzv': 0.00463528097789879,
  'ATSC3p': 0.004558142291516235,
  'ATSC2v': -0.004550088266698585,
  'ATSC5se': -0.00442243501281143,
  'ATSC0i': 0.004314492190282043,
  'AATS0i': 0.004241714262895952,
  'SMR_VSA4': 0.004214130514558937,
  'Mor12m': 0.004123668433162665,
  'GATS8Z': 0.003878508009245308,
  'GATS2pe': 0.0038411985668661656,
  'ATSC2se': -0.0036228142018597045,
  'Mor05v': -0.0034515466442078917,
  'BertzCT': 0.0033810458092780635,
  'GATS1d': 0.0033145741419369507,
  'AATSC7dv': 0.0032971353646910977,
  'Mor32m': -0.0031899890121089096,
  'GATS1pe': -0.0031599525391522725,
  'Mor04m': -0.002994585785813329,
  'Xpc-4d': 0.0029481044932680657,
  'ATSC5c': -0.0027035010160079047,
  'GATS7v': -0.0026707947283959163,
  'C1SP2': 0.002609641521495827,
  'BalabanJ': -0.0026005323726227417,
  'Mor28m': -0.002558739893057113,
  'Mor08p': 0.002402515626697463,
  'ATS0dv': 0.002391766968703715,
  'ATSC4p': 0.0023100954233423484,
  'AATS0s': -0.002283643270338704,
  'SlogP_VSA8': -0.0021781827981010656,
  'ATSC7d': -0.002157439015782142,
  'AATS7s': -0.0021031222183819803,
  'AATS4i': 0.0018362999805065171,
  'AATSC8Z': -0.0017249952237436636,
  'AATSC8s': -0.0013104399048625135,
  'ETA_eta': -0.0012892607379178737,
  'AATSC0v': 0.0011370541916987963,
  'ETA_beta_s': -0.0011002668124823179,
  'GGI1': -0.00102978149010937,
  'Xc-5d': 0.0009812545095996433,
  'ATS6Z': -0.000951268458762142,
  'AATS8i': -0.0006175520907965654,
  'ATS7Z': 0.0006012797445065009,
  'AATSC8d': 0.0005781758748009287,
  'GATS5i': -0.0005646279557973399,
  'Xc-3dv': 0.0005191797356694013,
  'ATSC5dv': -0.0004020036125952537}}

feature_order ensures that the trained model sees features in the exact same order as trained.

[9]:

feature_order[:5]

[9]:

['ABC', 'ABCGG', 'nBase', 'SpDiam_A', 'nAromAtom']

[10]:

from chemml.chem import Molecule
from sklearn.preprocessing import StandardScaler
from sklearn.model_selection import train_test_split
import tqdm
from chemml.chem import Mordred
from sklearn.svm import SVR

mol_objs_list = []
for i, smi in enumerate(tqdm.tqdm(df["smiles"], desc="Converting SMILES to ChemML Molecule objects")):
    mol = Molecule(smi, 'smiles')
    mol.hydrogens('add')
    try:
        mol.to_xyz('MMFF', maxIters=10000, mmffVariant='MMFF94s')
        mol_objs_list.append(mol)
    except Exception as e:
        print(f"\nUnable to process SMILES: {smi}; Error: {e}")
        # self.discarded_indices.append(i)
loaded_mols = mol_objs_list
features = Mordred(ignore_3D=False).represent(loaded_mols).drop(columns=['SMILES'])[feature_order].values
targets = df["density_class"].values


X_test = loaded_scaler.transform(features)

y_pred = loaded_model.predict(X_test)
y_pred_proba = loaded_model.predict_proba(X_test)

Converting SMILES to ChemML Molecule objects: 100%|██████████| 200/200 [00:06<00:00, 29.34it/s]

Step 1: Normalizing molecule input
Step 2/3: Optimizing geometry for 3D descriptor calculation
Step 3/3: Calculating Mordred descriptors

The loaded_model object can be treated just like any model object

[11]:

loaded_model

[11]:

LogisticRegression(C=0.1, solver='sag')

In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook.
On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.

LogisticRegression

?Documentation for LogisticRegressioniFitted

Parameters

	C C: float, default=1.0 Inverse of regularization strength; must be a positive float. Like in support vector machines, smaller values specify stronger regularization. `C=np.inf` results in unpenalized logistic regression. For a visual example on the effect of tuning the `C` parameter with an L1 penalty, see: :ref:`sphx_glr_auto_examples_linear_model_plot_logistic_path.py`.	0.1
	solver solver: {'lbfgs', 'liblinear', 'newton-cg', 'newton-cholesky', 'sag', 'saga'}, default='lbfgs' Algorithm to use in the optimization problem. Default is 'lbfgs'. To choose a solver, you might want to consider the following aspects: - 'lbfgs' is a good default solver because it works reasonably well for a wide class of problems. - For :term:`multiclass` problems (`n_classes >= 3`), all solvers except 'liblinear' minimize the full multinomial loss, 'liblinear' will raise an error. - 'newton-cholesky' is a good choice for `n_samples` >> `n_features * n_classes`, especially with one-hot encoded categorical features with rare categories. Be aware that the memory usage of this solver has a quadratic dependency on `n_features * n_classes` because it explicitly computes the full Hessian matrix. - For small datasets, 'liblinear' is a good choice, whereas 'sag' and 'saga' are faster for large ones; - 'liblinear' can only handle binary classification by default. To apply a one-versus-rest scheme for the multiclass setting one can wrap it with the :class:`~sklearn.multiclass.OneVsRestClassifier`. .. warning:: The choice of the algorithm depends on the penalty chosen (`l1_ratio=0` for L2-penalty, `l1_ratio=1` for L1-penalty and `0 < l1_ratio < 1` for Elastic-Net) and on (multinomial) multiclass support: ================= ======================== ====================== solver l1_ratio multinomial multiclass ================= ======================== ====================== 'lbfgs' l1_ratio=0 yes 'liblinear' l1_ratio=1 or l1_ratio=0 no 'newton-cg' l1_ratio=0 yes 'newton-cholesky' l1_ratio=0 yes 'sag' l1_ratio=0 yes 'saga' 0<=l1_ratio<=1 yes ================= ======================== ====================== .. note:: 'sag' and 'saga' fast convergence is only guaranteed on features with approximately the same scale. You can preprocess the data with a scaler from :mod:`sklearn.preprocessing`. .. seealso:: Refer to the :ref:`User Guide <Logistic_regression>` for more information regarding :class:`LogisticRegression` and more specifically the :ref:`Table <logistic_regression_solvers>` summarizing solver/penalty supports. .. versionadded:: 0.17 Stochastic Average Gradient (SAG) descent solver. Multinomial support in version 0.18. .. versionadded:: 0.19 SAGA solver. .. versionchanged:: 0.22 The default solver changed from 'liblinear' to 'lbfgs' in 0.22. .. versionadded:: 1.2 newton-cholesky solver. Multinomial support in version 1.6.	'sag'
	penalty penalty: {'l1', 'l2', 'elasticnet', None}, default='l2' Specify the norm of the penalty: - `None`: no penalty is added; - `'l2'`: add an L2 penalty term and it is the default choice; - `'l1'`: add an L1 penalty term; - `'elasticnet'`: both L1 and L2 penalty terms are added. .. warning:: Some penalties may not work with some solvers. See the parameter `solver` below, to know the compatibility between the penalty and solver. .. versionadded:: 0.19 l1 penalty with SAGA solver (allowing 'multinomial' + L1) .. deprecated:: 1.8 `penalty` was deprecated in version 1.8 and will be removed in 1.10. Use `l1_ratio` and `C` instead. `l1_ratio=0` for `penalty='l2'`, `l1_ratio=1` for `penalty='l1'`, `l1_ratio` set to any float between 0 and 1 for `penalty='elasticnet'`, and `C=np.inf` for `penalty=None`.	'deprecated'
	l1_ratio l1_ratio: float, default=0.0 The Elastic-Net mixing parameter, with `0 <= l1_ratio <= 1`. Setting `l1_ratio=1` gives a pure L1-penalty, setting `l1_ratio=0` a pure L2-penalty. Any value between 0 and 1 gives an Elastic-Net penalty of the form `l1_ratio * L1 + (1 - l1_ratio) * L2`. .. warning:: Certain values of `l1_ratio`, i.e. some penalties, may not work with some solvers. See the parameter `solver` below, to know the compatibility between the penalty and solver. .. versionchanged:: 1.8 Default value changed from None to 0.0. .. deprecated:: 1.8 `None` is deprecated and will be removed in version 1.10. Always use `l1_ratio` to specify the penalty type.	0.0
	dual dual: bool, default=False Dual (constrained) or primal (regularized, see also :ref:`this equation <regularized-logistic-loss>`) formulation. Dual formulation is only implemented for l2 penalty with liblinear solver. Prefer `dual=False` when n_samples > n_features.	False
	tol tol: float, default=1e-4 Tolerance for stopping criteria.	0.0001
	fit_intercept fit_intercept: bool, default=True Specifies if a constant (a.k.a. bias or intercept) should be added to the decision function.	True
	intercept_scaling intercept_scaling: float, default=1 Useful only when the solver `liblinear` is used and `self.fit_intercept` is set to `True`. In this case, `x` becomes `[x, self.intercept_scaling]`, i.e. a "synthetic" feature with constant value equal to `intercept_scaling` is appended to the instance vector. The intercept becomes ``intercept_scaling * synthetic_feature_weight``. .. note:: The synthetic feature weight is subject to L1 or L2 regularization as all other features. To lessen the effect of regularization on synthetic feature weight (and therefore on the intercept) `intercept_scaling` has to be increased.	1
	class_weight class_weight: dict or 'balanced', default=None Weights associated with classes in the form ``{class_label: weight}``. If not given, all classes are supposed to have weight one. The "balanced" mode uses the values of y to automatically adjust weights inversely proportional to class frequencies in the input data as ``n_samples / (n_classes * np.bincount(y))``. Note that these weights will be multiplied with sample_weight (passed through the fit method) if sample_weight is specified. .. versionadded:: 0.17 class_weight='balanced'	None
	random_state random_state: int, RandomState instance, default=None Used when ``solver`` == 'sag', 'saga' or 'liblinear' to shuffle the data. See :term:`Glossary <random_state>` for details.	None
	max_iter max_iter: int, default=100 Maximum number of iterations taken for the solvers to converge.	100
	verbose verbose: int, default=0 For the liblinear and lbfgs solvers set verbose to any positive number for verbosity.	0
	warm_start warm_start: bool, default=False When set to True, reuse the solution of the previous call to fit as initialization, otherwise, just erase the previous solution. Useless for liblinear solver. See :term:`the Glossary <warm_start>`. .. versionadded:: 0.17 warm_start to support lbfgs, newton-cg, sag, saga solvers.	False
	n_jobs n_jobs: int, default=None Does not have any effect. .. deprecated:: 1.8 `n_jobs` is deprecated in version 1.8 and will be removed in 1.10.	None

Fitted attributes

Name	Type	Value
classes_ classes_: ndarray of shape (n_classes, ) A list of class labels known to the classifier.	ndarray[int64](3,)	[0,1,2]
coef_ coef_: ndarray or CSR matrix of shape (1, n_features) or (n_classes, n_features) Coefficients of the features in the decision function. `coef_` is of shape (1, n_features) when the given problem is binary. By default, it will be created as a dense array, but can be turned to sparse (CSR format) through :meth:`sparsify` (which can be beneficial under L1 regularization when many coefficients are zero), and back to dense through :meth:`densify`.	ndarray[float64](3, 530)	[[ 0.02, 0.01,-0.02,..., 0.01,-0.06,-0.04], [-0. ,-0.02, 0. ,..., 0.03, 0.01, 0.01], [-0.01, 0.01, 0.02,...,-0.03, 0.05, 0.03]]
intercept_ intercept_: ndarray of shape (1,) or (n_classes,) Intercept (a.k.a. bias) added to the decision function. If `fit_intercept` is set to False, the intercept is set to zero. `intercept_` is of shape (1,) when the given problem is binary.	ndarray[float64](3,)	[-0.49, 0.87,-0.38]
n_features_in_ n_features_in_: int Number of features seen during :term:`fit`. .. versionadded:: 0.24	int	530
n_iter_ n_iter_: ndarray of shape (1, ) Actual number of iterations for all classes. .. versionchanged:: 0.20 In SciPy <= 1.0.0 the number of lbfgs iterations may exceed ``max_iter``. ``n_iter_`` will now report at most ``max_iter``.	ndarray[int32](1,)	[100]

[12]:

targets[:5], y_pred[:5]

[12]:

([2, 0, 2, 2, 0]
 Categories (3, int64): [0 < 1 < 2],
 array([2, 0, 2, 2, 0]))

[13]:

from chemml.utils import classification_metrics
classification_metrics(targets, y_pred)

TP:  [65 66 65]
FP:  [0 3 1]
TN:  [131 130 131]
FN : [2 4 2]
Accuracy:  0.98
Precision:  [1.         0.95652174 0.98484848]
Recall:  [0.97014925 1.         0.97014925]
f1_score:  [0.98484848 0.97777778 0.97744361]
FPR:  [0.         0.02255639 0.00757576]
TPR:  [0.97014925 0.94285714 0.97014925]
                    0           1           2
TP          65.000000   66.000000   65.000000
TN         131.000000  130.000000  131.000000
FP           0.000000    3.000000    1.000000
FN           2.000000    4.000000    2.000000
precision    1.000000    0.956522    0.984848
recall       0.970149    1.000000    0.970149
f1_score     0.984848    0.977778    0.977444
TPR          0.970149    0.942857    0.970149
FPR          0.000000    0.022556    0.007576
TNR          1.000000    0.977444    0.992424
PPV          1.000000    0.956522    0.984848
NPV          0.984962    0.970149    0.984962
FNR          0.029851    0.057143    0.029851
FDR          0.000000    0.043478    0.015152

[13]:

(np.float64(0.98),
       0     1     2
 0  65.0   1.0   1.0
 1   0.0  66.0   0.0
 2   0.0   2.0  65.0,
                     0           1           2
 TP          65.000000   66.000000   65.000000
 TN         131.000000  130.000000  131.000000
 FP           0.000000    3.000000    1.000000
 FN           2.000000    4.000000    2.000000
 precision    1.000000    0.956522    0.984848
 recall       0.970149    1.000000    0.970149
 f1_score     0.984848    0.977778    0.977444
 TPR          0.970149    0.942857    0.970149
 FPR          0.000000    0.022556    0.007576
 TNR          1.000000    0.977444    0.992424
 PPV          1.000000    0.956522    0.984848
 NPV          0.984962    0.970149    0.984962
 FNR          0.029851    0.057143    0.029851
 FDR          0.000000    0.043478    0.015152)

[14]:

%matplotlib inline
from chemml.visualization import ClassificationPlots


plotter = ClassificationPlots().plot(targets, y_pred, y_pred_proba)
display(plotter)

../_images/ipython_notebooks_autoML_classification_21_0.png