lorenzo.volpi
/
QuaPy
forked from moreo/QuaPy
1
0
Fork 0
Code Pull Requests Activity

mlq

This commit is contained in:
Alejandro Moreo Fernandez 2021-08-25 17:08:06 +02:00
parent 60b6fa3c12
commit c6de5a043d
2 changed files with 49 additions and 25 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

68
MultiLabel/multi_label.py → MultiLabel/main.py
View File

@ -2,7 +2,8 @@ from sklearn.calibration import CalibratedClassifierCV
from sklearn.linear_model import LogisticRegression from sklearn.linear_model import LogisticRegression
from sklearn.multioutput import ClassifierChain from sklearn.multioutput import ClassifierChain
from tqdm import tqdm from tqdm import tqdm
from skmultilearn.dataset import load_dataset
from scipy.sparse import csr_matrix
import quapy as qp import quapy as qp
from MultiLabel.mlclassification import MultilabelStackedClassifier from MultiLabel.mlclassification import MultilabelStackedClassifier
from MultiLabel.mldata import MultilabelledCollection from MultiLabel.mldata import MultilabelledCollection
@ -13,7 +14,7 @@ from method.aggregative import PACC, CC, EMQ, PCC, ACC, HDy
import numpy as np import numpy as np
from data.dataset import Dataset from data.dataset import Dataset
from mlevaluation import ml_natural_prevalence_evaluation, ml_artificial_prevalence_evaluation from mlevaluation import ml_natural_prevalence_evaluation, ml_artificial_prevalence_evaluation
import sys
def cls(): def cls():
# return LinearSVC() # return LinearSVC()
@ -31,24 +32,24 @@ n_samples = 5000
def models(): def models():
# yield 'NaiveCC', MultilabelNaiveAggregativeQuantifier(CC(cls())) yield 'NaiveCC', MultilabelNaiveAggregativeQuantifier(CC(cls()))
# yield 'NaivePCC', MultilabelNaiveAggregativeQuantifier(PCC(cls())) yield 'NaivePCC', MultilabelNaiveAggregativeQuantifier(PCC(cls()))
# yield 'NaiveACC', MultilabelNaiveAggregativeQuantifier(ACC(cls())) yield 'NaiveACC', MultilabelNaiveAggregativeQuantifier(ACC(cls()))
# yield 'NaivePACC', MultilabelNaiveAggregativeQuantifier(PACC(cls())) yield 'NaivePACC', MultilabelNaiveAggregativeQuantifier(PACC(cls()))
# yield 'EMQ', MultilabelQuantifier(EMQ(calibratedCls())) # yield 'EMQ', MultilabelQuantifier(EMQ(calibratedCls()))
# yield 'StackCC', MLCC(MultilabelStackedClassifier(cls())) yield 'StackCC', MLCC(MultilabelStackedClassifier(cls()))
# yield 'StackPCC', MLPCC(MultilabelStackedClassifier(cls())) yield 'StackPCC', MLPCC(MultilabelStackedClassifier(cls()))
# yield 'StackACC', MLACC(MultilabelStackedClassifier(cls())) yield 'StackACC', MLACC(MultilabelStackedClassifier(cls()))
# yield 'StackPACC', MLPACC(MultilabelStackedClassifier(cls())) yield 'StackPACC', MLPACC(MultilabelStackedClassifier(cls()))
# yield 'ChainCC', MLCC(ClassifierChain(cls(), cv=None, order='random')) # yield 'ChainCC', MLCC(ClassifierChain(cls(), cv=None, order='random'))
# yield 'ChainPCC', MLPCC(ClassifierChain(cls(), cv=None, order='random')) # yield 'ChainPCC', MLPCC(ClassifierChain(cls(), cv=None, order='random'))
# yield 'ChainACC', MLACC(ClassifierChain(cls(), cv=None, order='random')) # yield 'ChainACC', MLACC(ClassifierChain(cls(), cv=None, order='random'))
# yield 'ChainPACC', MLPACC(ClassifierChain(cls(), cv=None, order='random')) # yield 'ChainPACC', MLPACC(ClassifierChain(cls(), cv=None, order='random'))
common={'sample_size':sample_size, 'n_samples': n_samples, 'norm': True, 'means':False, 'stds':False, 'regression':'svr'} common={'sample_size':sample_size, 'n_samples': n_samples, 'norm': True, 'means':False, 'stds':False, 'regression':'svr'}
# yield 'MRQ-CC', MLRegressionQuantification(MultilabelNaiveQuantifier(CC(cls())), **common) yield 'MRQ-CC', MLRegressionQuantification(MultilabelNaiveQuantifier(CC(cls())), **common)
# yield 'MRQ-PCC', MLRegressionQuantification(MultilabelNaiveQuantifier(PCC(cls())), **common) yield 'MRQ-PCC', MLRegressionQuantification(MultilabelNaiveQuantifier(PCC(cls())), **common)
# yield 'MRQ-ACC', MLRegressionQuantification(MultilabelNaiveQuantifier(ACC(cls())), **common) yield 'MRQ-ACC', MLRegressionQuantification(MultilabelNaiveQuantifier(ACC(cls())), **common)
# yield 'MRQ-PACC', MLRegressionQuantification(MultilabelNaiveQuantifier(PACC(cls())), **common) yield 'MRQ-PACC', MLRegressionQuantification(MultilabelNaiveQuantifier(PACC(cls())), **common)
# yield 'MRQ-StackCC', MLRegressionQuantification(MLCC(MultilabelStackedClassifier(cls())), **common) # yield 'MRQ-StackCC', MLRegressionQuantification(MLCC(MultilabelStackedClassifier(cls())), **common)
# yield 'MRQ-StackPCC', MLRegressionQuantification(MLPCC(MultilabelStackedClassifier(cls())), **common) # yield 'MRQ-StackPCC', MLRegressionQuantification(MLPCC(MultilabelStackedClassifier(cls())), **common)
# yield 'MRQ-StackACC', MLRegressionQuantification(MLACC(MultilabelStackedClassifier(cls())), **common) # yield 'MRQ-StackACC', MLRegressionQuantification(MLACC(MultilabelStackedClassifier(cls())), **common)
@ -63,19 +64,36 @@ def models():
# yield 'MRQ-ChainPACC', MLRegressionQuantification(MLPACC(ClassifierChain(cls())), **common) # yield 'MRQ-ChainPACC', MLRegressionQuantification(MLPACC(ClassifierChain(cls())), **common)
dataset = 'reuters21578' # dataset = 'reuters21578'
picklepath = '/home/moreo/word-class-embeddings/pickles' # picklepath = '/home/moreo/word-class-embeddings/pickles'
data = Dataset.load(dataset, pickle_path=f'{picklepath}/{dataset}.pickle') # data = Dataset.load(dataset, pickle_path=f'{picklepath}/{dataset}.pickle')
# Xtr, Xte = data.vectorize()
Xtr, Xte = data.vectorize() # ytr = data.devel_labelmatrix.todense().getA()
ytr = data.devel_labelmatrix.todense().getA() # yte = data.test_labelmatrix.todense().getA()
yte = data.test_labelmatrix.todense().getA()
# remove categories with < 10 training documents # remove categories with < 10 training documents
to_keep = np.logical_and(ytr.sum(axis=0)>=50, yte.sum(axis=0)>=50) # to_keep = np.logical_and(ytr.sum(axis=0)>=50, yte.sum(axis=0)>=50)
ytr = ytr[:, to_keep] # ytr = ytr[:, to_keep]
yte = yte[:, to_keep] # yte = yte[:, to_keep]
print(f'num categories = {ytr.shape[1]}') # print(f'num categories = {ytr.shape[1]}')
dataset = 'birds'
Xtr, ytr, feature_names, label_names = load_dataset(dataset, 'train')
Xte, yte, _, _ = load_dataset(dataset, 'test')
print(f'n-labels = {len(label_names)}')
Xtr = csr_matrix(Xtr)
Xte = csr_matrix(Xte)
ytr = ytr.todense().getA()
yte = yte.todense().getA()
# print((np.abs(np.corrcoef(ytr, rowvar=False))>0.1).sum())
# sys.exit(0)
train = MultilabelledCollection(Xtr, ytr) train = MultilabelledCollection(Xtr, ytr)
test = MultilabelledCollection(Xte, yte) test = MultilabelledCollection(Xte, yte)

6
MultiLabel/mlquantification.py
View File

@ -186,6 +186,7 @@ class MLRegressionQuantification:
# self.norm = StandardScaler() # self.norm = StandardScaler()
self.means = means self.means = means
self.stds = stds self.stds = stds
# self.covs = covs
def _prepare_arrays(self, Xs, ys, samples_mean, samples_std): def _prepare_arrays(self, Xs, ys, samples_mean, samples_std):
Xs = np.asarray(Xs) Xs = np.asarray(Xs)
@ -196,6 +197,8 @@ class MLRegressionQuantification:
if self.stds: if self.stds:
samples_std = np.asarray(samples_std) samples_std = np.asarray(samples_std)
Xs = np.hstack([Xs, samples_std]) Xs = np.hstack([Xs, samples_std])
# if self.covs:
return Xs, ys return Xs, ys
def generate_samples_npp(self, val): def generate_samples_npp(self, val):
@ -257,3 +260,6 @@ class MLRegressionQuantification:
adjusted = adjusted.flatten() adjusted = adjusted.flatten()
neg_prevs = 1-adjusted neg_prevs = 1-adjusted
return np.asarray([neg_prevs, adjusted]).T return np.asarray([neg_prevs, adjusted]).T
# class