Python-EEG工具库MNE中文教程(5)-机器学习算法随机森林判断睡眠类型

import numpy as np
import matplotlib.pyplot as plt

import mne
from mne.datasets.sleep_physionet.age import fetch_data
from mne.time_frequency import psd_welch

from sklearn.ensemble import RandomForestClassifier
from sklearn.metrics import accuracy_score
from sklearn.metrics import confusion_matrix
from sklearn.metrics import classification_report
from sklearn.pipeline import make_pipeline
from sklearn.preprocessing import FunctionTransformer

"""
可以通过
mne.datasets.sleep_physionet.age.fetch_data(subjects,recording,path)
来获取PhysioNet多导睡眠图数据集文件。
subjects：表示想要使用哪些受试者对象，可供选择的受试者对象范围为0-19。
recording：表示夜间记录的编号(索引)，有效值为：[1]、[2]或[1、2]。

path:PhysioNet数据的存放地址，如果没有给定，则加载默认存放数据的地址；
如果默认存放数据集的地址不存在数据，则从网络中下载相关数据。
"""
# 选择两个受试者实验对象ALICE, BOB(该名字并非实验中的真实名，这里是为了方便才临时取的名字)
ALICE, BOB = 0, 1
# 加载ALICE, BOB的实验数据文件
[alice_files, bob_files] = fetch_data(subjects=[ALICE, BOB], recording=[1])

# 通道名称映射
mapping = {'EOG horizontal': 'eog',
           'Resp oro-nasal': 'misc',
           'EMG submental': 'misc',
           'Temp rectal': 'misc',
           'Event marker': 'misc'}

#读取ALICE的edf文件，和其对应的注释文件
raw_train = mne.io.read_raw_edf(alice_files[0])
annot_train = mne.read_annotations(alice_files[1])

raw_train.set_annotations(annot_train, emit_warning=False)
raw_train.set_channel_types(mapping)

# 绘制空0s开始，时间窗口长度为40s的连续通道数据波形图
raw_train.plot(duration=40, scalings='auto')
plt.show()

"""
睡眠表示与事件映射
"""
annotation_desc_2_event_id = {'Sleep stage W': 1,
                              'Sleep stage 1': 2,
                              'Sleep stage 2': 3,
                              'Sleep stage 3': 4,
                              'Sleep stage 4': 4,
                              'Sleep stage R': 5}

events_train, _ = mne.events_from_annotations(
    raw_train, event_id=annotation_desc_2_event_id, chunk_duration=30.)

# 创建一个新的event_id以统一 阶段3和4
event_id = {'Sleep stage W': 1,
            'Sleep stage 1': 2,
            'Sleep stage 2': 3,
            'Sleep stage 3/4': 4,
            'Sleep stage R': 5}

# 绘制事件数据
mne.viz.plot_events(events_train, event_id=event_id,
                    sfreq=raw_train.info['sfreq'])

# 保留颜色代码以便进一步绘制
stage_colors = plt.rcParams['axes.prop_cycle'].by_key()['color']

tmax = 30. - 1. / raw_train.info['sfreq']  # tmax in included
"""
所创建的是时间从tmin=0开始，到tmax为止的epochs
"""
epochs_train = mne.Epochs(raw=raw_train, events=events_train,
                          event_id=event_id, tmin=0., tmax=tmax, baseline=None)

print(epochs_train)

raw_test = mne.io.read_raw_edf(bob_files[0])
annot_test = mne.read_annotations(bob_files[1])
raw_test.set_annotations(annot_test, emit_warning=False)
raw_test.set_channel_types(mapping)
events_test, _ = mne.events_from_annotations(
    raw_test, event_id=annotation_desc_2_event_id, chunk_duration=30.)
epochs_test = mne.Epochs(raw=raw_test, events=events_test, event_id=event_id,
                         tmin=0., tmax=tmax, baseline=None)

print(epochs_test)

fig, (ax1, ax2) = plt.subplots(ncols=2)

# iterate over the subjects
stages = sorted(event_id.keys())
for ax, title, epochs in zip([ax1, ax2],
                             ['Alice', 'Bob'],
                             [epochs_train, epochs_test]):

    for stage, color in zip(stages, stage_colors):
        epochs[stage].plot_psd(area_mode=None, color=color, ax=ax,
                               fmin=0.1, fmax=20., show=False,
                               average=True, spatial_colors=False)
    ax.set(title=title, xlabel='Frequency (Hz)')
ax2.set(ylabel='uV^2/hz (dB)')
ax2.legend(ax2.lines[2::3], stages)
plt.tight_layout()
plt.show()

def eeg_power_band(epochs):
    """脑电相对功率带特征提取
    该函数接受一个""mne.Epochs"对象，
    并基于与scikit-learn兼容的特定频带中的相对功率创建EEG特征。
    Parameters
    ----------
    epochs : Epochs
        The data.
    Returns
    -------
    X : numpy array of shape [n_samples, 5]
        Transformed data.
    """
    # 特定频带
    FREQ_BANDS = {"delta": [0.5, 4.5],
                  "theta": [4.5, 8.5],
                  "alpha": [8.5, 11.5],
                  "sigma": [11.5, 15.5],
                  "beta": [15.5, 30]}

    psds, freqs = psd_welch(epochs, picks='eeg', fmin=0.5, fmax=30.)
    # 归一化 PSDs
    psds /= np.sum(psds, axis=-1, keepdims=True)

    X = []
    for fmin, fmax in FREQ_BANDS.values():
        psds_band = psds[:, :, (freqs >= fmin) & (freqs < fmax)].mean(axis=-1)
        X.append(psds_band.reshape(len(psds), -1))

    return np.concatenate(X, axis=1)

pipe = make_pipeline(FunctionTransformer(eeg_power_band, validate=False),
                     RandomForestClassifier(n_estimators=100, random_state=42))
# 训练
y_train = epochs_train.events[:, 2]
pipe.fit(epochs_train, y_train)

# 预测
y_pred = pipe.predict(epochs_test)

# 评估准确率
y_test = epochs_test.events[:, 2]
acc = accuracy_score(y_test, y_pred)

print("Accuracy score: {}".format(acc))

print(classification_report(y_test, y_pred, target_names=event_id.keys()))