aaanalysis.TreeModel.predict_proba

TreeModel.predict_proba(X)[source]

Obtain Monte Carlo estimate of class prediction probabilities for the positive class in X.

Predictions are performed using all tree-based models from the list_model_classes attribute and feature selections from the is_selected attribute.

Parameters:

X (array-like, shape (n_samples, n_features)) – Feature matrix. Rows typically correspond to samples and columns to features.

Returns:

pred (array-like, shape (n_samples)) – Array with the average prediction score for the positive class for each sample.
pred_std (array-like, shape (n_samples)) – Array with the standard deviation of prediction scores for the positive class for each sample.

See also

sklearn.ensemble.RandomForestClassifier for information on tree-based models.
TreeModel.fit() for fitting the TreeModel object.

Examples

To demonstrate the TreeModel().predict_proba()method, we obtain the DOM_GSEC example dataset and its respective feature set (see [Breimann25a]):

import aaanalysis as aa
aa.options["verbose"] = False # Disable verbosity

df_seq = aa.load_dataset(name="DOM_GSEC")
labels = df_seq["label"].to_list()
df_feat = aa.load_features(name="DOM_GSEC").head(100)

# Create feature matrix
sf = aa.SequenceFeature()
df_parts = sf.get_df_parts(df_seq=df_seq)
X = sf.feature_matrix(features=df_feat["feature"], df_parts=df_parts)

We can not fit the TreeModel, which will internally fit 3 tree-based models over 5 training rounds be default:

tm = aa.TreeModel()
tm = tm.fit(X, labels=labels)

Using the TreeModel().predict_proba() method calculates probability predictions by averaging across multiple models and rounds, using a Monte Carlo approach for robust estimation:

pred, pred_std = tm.predict_proba(X)

df_seq["prediction"] = pred
df_seq["pred_std"] = pred_std

print("Prediction scores for 5 substrates")
aa.display_df(df_seq.head(5))

print("Prediction scores for 5 non-substrates")
aa.display_df(df_seq.tail(5))

Prediction scores for 5 substrates

	entry	sequence	label	tmd_start	tmd_stop	jmd_n	tmd	jmd_c	prediction	pred_std
1	P05067	MLPGLALLLLAAWTA...GYENPTYKFFEQMQN	1	701	723	FAEDVGSNKG	AIIGLMVGGVVIATVIVITLVML	KKKQYTSIHH	0.981000	0.013928
2	P14925	MAGRARSGLLLLLLG...EEEYSAPLPKPAPSS	1	868	890	KLSTEPGSGV	SVVLITTLLVIPVLVLLAIVMFI	RWKKSRAFGD	0.973000	0.012884
3	P70180	MRSLLLFTFSACVLL...RELREDSIRSHFSVA	1	477	499	PCKSSGGLEE	SAVTGIVVGALLGAGLLMAFYFF	RKKYRITIER	0.976000	0.008000
4	Q03157	MGPTSPAARGQGRRW...HGYENPTYRFLEERP	1	585	607	APSGTGVSRE	ALSGLLIMGAGGGSLIVLSLLLL	RKKKPYGTIS	0.989000	0.008000
5	Q06481	MAATGTAAAAATGRL...GYENPTYKYLEQMQI	1	694	716	LREDFSLSSS	ALIGLLVIAVAIATVIVISLVML	RKRQYGTISH	0.996000	0.003742

Prediction scores for 5 non-substrates

	entry	sequence	tmd_start	tmd_stop	jmd_n	tmd	jmd_c	prediction	pred_std
122	P36941	MLLPWATSAPGLAWG...TPSNRGPRNQFITHD	226	248	PLPPEMSGTM	LMLAVLLPLAFFLLLATVFSCIW	KSHPSLCRKL	0.008000	0.005099
123	P25446	MLWIWAVLPLVLAGS...STPDTGNENEGQCLE	170	187	NCRKQSPRNR	LWLLTILVLLIPLVFIYR	KYRKRKCWKR	0.150000	0.025884
124	Q9P2J2	MVWCLGLAVLSLVIS...AYRQPVPHPEQATLL	738	760	PGLLPQPVLA	GVVGGVCFLGVAVLVSILAGCLL	NRRRAARRRR	0.122000	0.017205
125	Q96J42	MVPAAGRRPPRVMRL...SIRWLIPGQEQEHVE	324	342	LPSTLIKSVD	WLLVFSLFFLISFIMYATI	RTESIRWLIP	0.037000	0.012490
126	P0DPA2	MRVGGAFHLLLVCLS...DCAEGPVQCKNGLLV	265	287	KVSDSRRIGV	IIGIVLGSLLALGCLAVGIWGLV	CCCCGGSGAG	0.011000	0.004899