- NLP
- nltk
- jieba
- 中文分词
- cut()
-
import jieba
mytext=" ".join(jieba.cut(mytext))
- boson NLP
- snownlp
- 用评论数据训练的模型,中文情感分析
- SnowNLP()z
- textblob
- NLTK
- Gensim
- spaCy
- 函数
- 对象
- displacy
- 可视化
- render
- render(doc, style=‘ent’, jupyter=True)
- render(doc, style=‘dep’, jupyter=True, options={‘distance’: 90})
- nlp
- pyldavis
- scikit-learn(sklearn)
- tsne
- 高维度向量压缩到二维平面
- fit_transform()
- 对象
- preprocessing
- LabelEncoder()
- OneHotEncoder()
- StandardScaler()
- 标准化
- fit_transform()
- transform()
- cross_validation
- train_test_split()
- cross_val_score()
- tree
- DecisionTreeClassifier(max_depth=3)
- metrics
- accuracy_scoreaccuracy_score()
- accuracy_score()
- confusion_matrix()
- classification_report()
- model_selection
- feature_extraction
- text
- CountVectorizer()
- 向量化处理
- 参数
- stop_words=
- max_df=0.8
- min_df=3
- token_pattern=u”
- fit_transform()
- get_feature_names()
- naive_bayes
- pipeline
- feature_extraction
- TfidfVectorizer()
- CountVectorizer()
- decomposition
- LatentDirichletAllocation()
- 参数
- n_topics=5
- max_iter=50
- learning_method=‘online’
- learning_offset=50.0
- random_state=0
- fit()
- 机器学习
- 深度学习
- TensorFlow
- 命令
- tensorboard —logdir=/tflearn_logs/
- feature_column
- numeric_column()
- categorical_column_with_vocabulary_list()
- indicator_column()
- data
- PyTorch
- Theano
- CNTK
- MXNet
- TFLearn
- 基于TensorFlow
- input_data()
- fully_connected(net, 6, activation=‘relu’)
- 添加一层隐藏层, 每层6个神经元
- 参数
- activation=‘relu’
- relu, 激活函数为ReLU
- softmax,输出层激活函数,可以表示第一分类的可能性
- regression()
- DNN()
- 生成模型
- fit()
- 开始拟合
- 参数
- n_epoch 训练轮次
- batch_size 每一次输入模型的数据行数
- show_metric 要不要打印过程
- predict()
- evaluate()
- Keras
- 基于TensorFlow
- Sequential()
- 对象
- utils
- preprocessing
- text
- Tokenizer()
- 词变数字
- 属性
- fit_on_texts()
- texts_to_sequences()
- sequence
- models
- Sequential()
- 构建顺序模型
- add()
- summary()
- layers
- 层list
- trainable
- set_weights()
- compile()
- 入参
- optimizer=‘rmsprop’
- loss=‘binary_crossentropy’
- metrics=[‘acc’]
- fit()
- 训练
- 入参
- epochs
- batch_size
- validation_data
- history
- save()
- layers
- Embedding()
- Flatten()
- Dense()
- DenseFeatures()
- LSTM()
- 用于add
- 参数
- dropout=0.2
- recurrent_dropout=0.2
- TensorLayer
- Turi Create
- TuriCreate
- 苹果开发用于移动设备赋能
- create()
- 开始训练
- predict()
- evaluate()
- Image()
- 对象
- image_analysis
- load_images()
- save()
- explore()
- random_split(0.8, seed=2)
- add_row_number()
- image_similarity