Python：BDOC算法

 Binary Decomposition ordinal classification

import numpy as np
import pandas as pd
from copy import deepcopy
from collections import OrderedDict
from sklearn.linear_model import LogisticRegression
from sklearn.metrics import accuracy_score, f1_score
from sklearn.model_selection import StratifiedKFold

def BDOC(X, y, absLabeled, test_idx):
    modelDict = OrderedDict()
    trainIndex = OrderedDict()
    trainTarget = OrderedDict()
    labels = np.unique(y)
    target = np.array([_ for _ in np.arange(labels[0],labels[-1],1)])
    ###----------------------------------
    for tar in target:
        modelDict[tar] = LogisticRegression(solver='newton-cg', penalty='l2')
        trainIndex[tar] = deepcopy(absLabeled)
        trainTarget[tar] = deepcopy(y[absLabeled])
        for j in range(len(absLabeled)):
            if trainTarget[tar][j] <= tar:
                trainTarget[tar][j] = tar
            else:
                trainTarget[tar][j] = tar + 1
    ###----------------------------------
    ###---------------------------------
    for tar, model in modelDict.items():
        model.fit(X[trainIndex[tar]], trainTarget[tar])
    probDict = OrderedDict()
    for tar, model in modelDict.items():
        probDict[tar] = model.predict_proba(X[test_idx])

    y_pred = []
    for j, idx in enumerate(test_idx):
        prob = OrderedDict()
        for ele in labels:
            if ele == labels[0]:
                prob[ele] = probDict[ele][j][0]
            elif ele == labels[-1]:
                prob[ele] = probDict[ele-1][j][1]
            else:
                prob[ele] = probDict[ele-1][j][1] - probDict[ele][j][1]
        y_pred.append(max(prob,key=prob.get))
    y_pred = np.array(y_pred)
    Acc = accuracy_score(y[test_idx],y_pred)
    return Acc

已验证！