主机屋网站在那注册,展览会建设网站平台的作用,如何设置网站布局,西安app开发软件定制关于 情绪检测#xff0c;是脑科学研究中的一个常见和热门的方向。在进行情绪检测的分类中#xff0c;真实数据不足#xff0c;经常导致情绪检测模型的性能不佳。因此#xff0c;对数据进行增强#xff0c;成为了一个提升下游任务的重要的手段。本项目通过DCGAN模型实现脑…关于 情绪检测是脑科学研究中的一个常见和热门的方向。在进行情绪检测的分类中真实数据不足经常导致情绪检测模型的性能不佳。因此对数据进行增强成为了一个提升下游任务的重要的手段。本项目通过DCGAN模型实现脑电信号的扩充。 图片来源https://www.medicalnewstoday.com/articles/seizure-eeg 工具 数据 方法实现 DCGAN速递https://arxiv.org/abs/1511.06434 数据加载和预处理 import numpy as np # linear algebra import pandas as pd # data processing, CSV file I/O (e.g. pd.read_csv) from sklearn.preprocessing import StandardScaler from tensorflow.keras.utils import to_categorical from sklearn.model_selection import train_test_split import tensorflow as tf from tensorflow.keras import Sequential from tensorflow.keras.layers import Dense, Dropout from tensorflow.keras.layers import Embedding from tensorflow.keras.layers import LSTM from tensorflow.keras.optimizers import SGD from sklearn.metrics import accuracy_score from model_DCGAN import DCGAN from tensorflow.keras.optimizers import Adam, SGD, RMSprop from sklearn.utils import shuffle from sklearn.ensemble import GradientBoostingClassifieruse_feature_reduction Truetf.keras.backend.clear_session()dfpd.read_csv(dataset/emotions.csv)encode ({NEUTRAL: 0, POSITIVE: 1, NEGATIVE: 2} ) #new dataset with replaced values df_encoded df.replace(encode)print(df_encoded.head()) print(df_encoded[label].value_counts()),xdf_encoded.drop([label] ,axis1) y df_encoded.loc[:,label].valuesscaler StandardScaler() scaler.fit(x) x scaler.transform(x) y to_categorical(y)x_train, x_test, y_train, y_test train_test_split(x, y, test_size 0.2, random_state 4)if use_feature_reduction:# Feature reduction partest GradientBoostingClassifier(n_estimators10, learning_rate0.1, random_state0).fit(x_train,y_train.argmax(-1))# Obtain feature importance results from Gradient Boosting Regressorfeature_importance est.feature_importances_epsilon_feature 1e-2x_train x_train[:, feature_importance epsilon_feature]x_test x_test[:, feature_importance epsilon_feature] 设置DCGAN优化器 # setup optimzers gen_optim Adam(1e-4, beta_10.5) disc_optim RMSprop(5e-4)训练GAN生成类别0脑电数据 # generate samples for class 0 generator_class 0 dcgan DCGAN(gen_optim, disc_optim, noise_dim100, dropout0.3, input_dimx_train.shape[2]) x_train_class_0 x_train[y_train[:,generator_class]1,:] loss_history_class_0, acc_history_class_0, grads_history_class_0 dcgan.train(x_train_class_0, epochs100) print(Class 0 fake samples are generating) generator_class_0 dcgan.generator generated_samples_class_0, _ dcgan.generate_fake_data(Nlen(x_train_class_0))训练GAN生成类别1脑电数据 # generate samples for class 1 generator_class 1 dcgan DCGAN(gen_optim, disc_optim, noise_dim100, dropout0.3, input_dimx_train.shape[2]) x_train_class_1 x_train[y_train[:,generator_class]1,:] loss_history_class_1, acc_history_class_1, grads_history_class_1 dcgan.train(x_train_class_1, epochs100) print(Class 1 fake samples are generating) generator_class_1 dcgan.generator generated_samples_class_1, _ dcgan.generate_fake_data(Nlen(x_train_class_1)) 训练GAN生成类别2脑电数据 # generate samples for class 2 generator_class 2 dcgan DCGAN(gen_optim, disc_optim, noise_dim100, dropout0.3, input_dimx_train.shape[2]) x_train_class_2 x_train[y_train[:,generator_class]1,:] loss_history_class_2, acc_history_class_2, grads_history_class_2 dcgan.train(x_train_class_2,epochs100) print(Class 2 fake samples are generating) generator_class_2 dcgan.generator generated_samples_class_2, _ dcgan.generate_fake_data(Nlen(x_train_class_2)) 合成数据融入真实训练数据集 generated_samples np.concatenate((generated_samples_class_0,generated_samples_class_1,generated_samples_class_2),axis0) generated_y np.concatenate((np.zeros((len(x_train_class_0),),dtypenp.int32),np.ones((len(x_train_class_1),),dtypenp.int32),2 * np.ones((len(x_train_class_2),),dtypenp.int32)),axis0)generated_y to_categorical(generated_y)x_train_all np.concatenate((x_train,generated_samples),axis0) y_train_all np.concatenate((y_train,generated_y), axis0)#shuffle training data x_train_all, y_train_all shuffle(x_train_all,y_train_all) 基于数据增强的LSTM模型情绪检测 model Sequential() model.add(LSTM(64, input_shape(1,x_train_all.shape[2]),activationrelu,return_sequencesTrue)) model.add(Dropout(0.2)) model.add(LSTM(32,activationsigmoid)) model.add(Dropout(0.2)) model.add(Dense(3, activationsigmoid)) model.compile(loss categorical_crossentropy, optimizer adam, metrics [accuracy]) model.summary()history model.fit(x_train_all, y_train_all, epochs 250, validation_data (x_test, y_test)) score, acc model.evaluate(x_test, y_test)pred model.predict(x_test) predict_classes np.argmax(pred,axis1) expected_classes np.argmax(y_test,axis1) print(expected_classes.shape) print(predict_classes.shape) correct accuracy_score(expected_classes,predict_classes) print(fTest Accuracy: {correct})已附DCGAN模型 相关项目和代码问题欢迎沟通交流。