摘要情绪识别作为自然语言处理NLP领域的重要分支在人机交互、社交媒体分析、客户服务等场景中具有广泛应用。本文系统介绍基于Python的情绪识别模型构建方法涵盖数据预处理、特征提取、模型选择、训练评估及部署应用等关键环节。通过完整的代码示例帮助读者快速掌握情绪识别模型的开发流程。1. 引言情绪识别旨在从文本、语音、图像等模态中自动识别人类情绪状态。文本情绪识别因其数据易获取、应用场景丰富而备受关注。Python凭借丰富的NLP生态如NLTK、Transformers、scikit-learn成为实现情绪识别的首选语言。本文聚焦文本情绪识别构建可识别愤怒、快乐、悲伤、恐惧、惊讶、厌恶等基本情绪的模型。2. 技术路线概览text数据采集 → 文本清洗 → 特征提取 → 模型训练 → 评估调优 → 部署应用3. 数据准备3.1 常用数据集数据集语言类别规模ISEAR英文7类约7600条GoEmotions英文27类5.8万条中文情感语料库中文6类约1.2万条ChnSentiCorp中文2-4类1万条3.2 数据加载示例pythonimport pandas as pd from sklearn.model_selection import train_test_split # 加载数据集示例数据结构text, label data pd.read_csv(emotion_dataset.csv) # label映射0-愤怒1-快乐2-悲伤3-恐惧4-惊讶5-厌恶 print(data[label].value_counts()) # 划分训练集和测试集 train_texts, test_texts, train_labels, test_labels train_test_split( data[text], data[label], test_size0.2, random_state42, stratifydata[label] )4. 文本预处理pythonimport re import jieba # 中文分词 from nltk.corpus import stopwords def preprocess_chinese(text): # 1. 去除特殊符号、数字、英文根据需求保留 text re.sub(r[a-zA-Z0-9], , text) text re.sub(r[^\u4e00-\u9fa5], , text) # 仅保留中文 # 2. 分词 words jieba.lcut(text) # 3. 去停用词 stop_words set(stopwords.words(chinese)) words [w for w in words if w not in stop_words and len(w) 1] return .join(words) # 应用预处理 train_texts_clean [preprocess_chinese(t) for t in train_texts]5. 特征提取方法5.1 TF-IDF传统方法pythonfrom sklearn.feature_extraction.text import TfidfVectorizer tfidf TfidfVectorizer(max_features5000, ngram_range(1,2)) X_train_tfidf tfidf.fit_transform(train_texts_clean) X_test_tfidf tfidf.transform(test_texts_clean)5.2 Word2Vec 平均池化pythonfrom gensim.models import Word2Vec import numpy as np # 训练Word2Vec或加载预训练模型 sentences [text.split() for text in train_texts_clean] w2v_model Word2Vec(sentences, vector_size128, window5, min_count2, workers4) def text_to_vector(text, model, vector_size): words text.split() vectors [model.wv[w] for w in words if w in model.wv] return np.mean(vectors, axis0) if vectors else np.zeros(vector_size) X_train_w2v np.array([text_to_vector(t, w2v_model, 128) for t in train_texts_clean])5.3 预训练模型BERT等pythonfrom transformers import AutoTokenizer, AutoModel import torch tokenizer AutoTokenizer.from_pretrained(bert-base-chinese) bert_model AutoModel.from_pretrained(bert-base-chinese) def get_bert_embedding(text): inputs tokenizer(text, return_tensorspt, truncationTrue, max_length128, paddingTrue) with torch.no_grad(): outputs bert_model(**inputs) return outputs.last_hidden_state[:, 0, :].numpy() # CLS token6. 模型构建与训练6.1 传统机器学习模型逻辑回归pythonfrom sklearn.linear_model import LogisticRegression from sklearn.metrics import classification_report, accuracy_score # 使用TF-IDF特征 lr_model LogisticRegression(max_iter1000, multi_classovr) lr_model.fit(X_train_tfidf, train_labels) y_pred lr_model.predict(X_test_tfidf) print(fAccuracy: {accuracy_score(test_labels, y_pred):.4f}) print(classification_report(test_labels, y_pred, target_names[愤怒,快乐,悲伤,恐惧,惊讶,厌恶]))6.2 深度学习模型TextCNNpythonimport torch.nn as nn import torch.nn.functional as F class TextCNN(nn.Module): def __init__(self, vocab_size, embed_dim, num_classes, filter_sizes[2,3,4], num_filters100): super(TextCNN, self).__init__() self.embedding nn.Embedding(vocab_size, embed_dim) self.convs nn.ModuleList([ nn.Conv2d(1, num_filters, (k, embed_dim)) for k in filter_sizes ]) self.dropout nn.Dropout(0.5) self.fc nn.Linear(len(filter_sizes) * num_filters, num_classes) def forward(self, x): x self.embedding(x) # (batch, seq_len, embed_dim) x x.unsqueeze(1) # (batch, 1, seq_len, embed_dim) conv_outs [] for conv in self.convs: conv_out F.relu(conv(x)).squeeze(3) # (batch, num_filters, seq_len - k 1) pool_out F.max_pool1d(conv_out, conv_out.size(2)).squeeze(2) conv_outs.append(pool_out) x torch.cat(conv_outs, dim1) x self.dropout(x) return self.fc(x) # 训练代码需配合DataLoader使用此处省略详细实现6.3 微调BERT推荐方案pythonfrom transformers import BertForSequenceClassification, Trainer, TrainingArguments model BertForSequenceClassification.from_pretrained(bert-base-chinese, num_labels6) training_args TrainingArguments( output_dir./results, num_train_epochs3, per_device_train_batch_size16, per_device_eval_batch_size64, warmup_steps500, weight_decay0.01, logging_dir./logs, evaluation_strategyepoch, ) # 需要将文本转换为BERT输入格式 def tokenize_function(texts): return tokenizer(texts, paddingTrue, truncationTrue, max_length128) train_encodings tokenize_function(train_texts_clean) test_encodings tokenize_function(test_texts_clean) # 创建Dataset类代码略 # trainer Trainer(...)7. 模型评估7.1 评估指标pythonfrom sklearn.metrics import confusion_matrix, ConfusionMatrixDisplay import matplotlib.pyplot as plt # 计算混淆矩阵 cm confusion_matrix(test_labels, y_pred) disp ConfusionMatrixDisplay(confusion_matrixcm, display_labels[愤怒,快乐,悲伤,恐惧,惊讶,厌恶]) disp.plot(cmapBlues) plt.show()7.2 各类别性能分析典型的情绪识别模型性能BERT微调后情绪类别精确率召回率F1分数愤怒0.850.830.84快乐0.900.920.91悲伤0.820.840.83恐惧0.780.750.76惊讶0.810.790.80厌恶0.790.770.788. 模型部署8.1 使用FastAPI部署pythonfrom fastapi import FastAPI, HTTPException from pydantic import BaseModel import joblib app FastAPI() # 加载模型和向量化器 model joblib.load(emotion_model.pkl) vectorizer joblib.load(tfidf_vectorizer.pkl) label_map {0: 愤怒, 1: 快乐, 2: 悲伤, 3: 恐惧, 4: 惊讶, 5: 厌恶} class TextRequest(BaseModel): text: str app.post(/predict) async def predict_emotion(request: TextRequest): # 预处理 cleaned preprocess_chinese(request.text) # 特征提取 features vectorizer.transform([cleaned]) # 预测 pred_label model.predict(features)[0] pred_proba model.predict_proba(features)[0].max() return { text: request.text, emotion: label_map[pred_label], confidence: float(pred_proba) } app.get(/health) async def health_check(): return {status: ok} # 启动命令uvicorn main:app --reload8.2 测试APIbashcurl -X POST http://localhost:8000/predict \ -H Content-Type: application/json \ -d {text:今天天气真好心情特别愉快}响应示例json{ text: 今天天气真好心情特别愉快, emotion: 快乐, confidence: 0.96 }9. 优化建议优化方向具体方法预期提升数据增强回译、同义词替换、随机删除2~5% F1特征融合TF-IDF BERT Embedding1~3% Acc模型集成投票/加权融合多模型3~6% Acc后处理阈值调整、序列平滑提升稳定性类别不平衡Focal Loss、过采样/欠采样改善少数类10. 完整项目结构textemotion_recognition/ ├── data/ │ ├── raw/ │ └── processed/ ├── models/ │ ├── bert_emotion/ │ └── saved_models/ ├── notebooks/ │ └── exploration.ipynb ├── src/ │ ├── preprocess.py │ ├── features.py │ ├── train.py │ ├── evaluate.py │ └── predict.py ├── api/ │ └── main.py ├── requirements.txt └── README.md11. 总结与展望本文系统介绍了基于Python的情绪识别模型开发全流程。从数据预处理到模型部署涵盖了传统机器学习和深度学习方法。实际应用中建议优先选择预训练模型如BERT、RoBERTa、ERNIE进行微调通常能获得最佳效果。未来方向多模态情绪识别融合文本、语音、面部表情细粒度情绪分析识别更复杂的复合情绪对话情绪追踪结合上下文理解情绪演变低资源场景少样本学习和跨语言迁移参考文献Devlin J, et al. BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding. NAACL 2019.Kim Y. Convolutional Neural Networks for Sentence Classification. EMNLP 2014.Demszky D, et al. GoEmotions: A Dataset of Fine-Grained Emotions. ACL 2020.