Course materials for Applied Natural Language Processing (Spring 2019). Syllabus: http://people.ischool.berkeley.edu/~dbamman/info256.html
| Notebook | Description |
|---|---|
| 1.words/EvaluateTokenizationForSentiment.ipynb | The impact of tokenization choices on sentiment classification. |
| 1.words/ExploreTokenization.ipynb | Different methods for tokenizing texts (whitespace, NLTK, spacy, regex) |
| 1.words/TokenizePrintedBooks.ipynb | Design a better tokenizer for printed books |
| 2.distinctive_terms/ChiSquare.ipynb | Find distinctive terms using the Chi-square test |
| 2.distinctive_terms/CompareCorpora.ipynb | Find distinctive terms using the Mann-Whitney rank sums test |
| 3.dictionaries/DictionaryTimeSeries.ipynb | Plot sentiment over time using human-defined dictionaries |
| 4.classification/CheckData_TODO.ipynb | Gather data for classification |
| 4.classification/FeatureExploration_TODO.ipynb | Feature engineering for text classification |
| 4.classification/FeatureWeights_TODO.ipynb | Analyze feature weights for text classification |
| 4.classification/Hyperparameters_TODO.ipynb | Explore hyperparameter choices on classification accuracy |
| 5.text_regression/Regularization.ipynb | Linear regression with L1/L2 regularization for box office prediction |
| 6.tests/BootstrapConfidenceIntervals.ipynb | Estimate confidence intervals with the bootstrap |
| 6.tests/ParametricTest.ipynb | Hypothesis testing with parametric (normal) tests |
| 6.tests/PermutationTest.ipynb | Hypothesis testing with non-parametric (permutation) tests |
| 7.embeddings/DistributionalSimilarity.ipynb | Explore distributional hypothesis to build high-dimensional, sparse representations for words |
| 7.embeddings/TFIDF.ipynb | Explore distributional hypothesis to build high-dimensional, sparse representations for words (with TF IDF scaling) |
| 7.embeddings/TurneyLittman2003.ipynb | Use word embeddings to implement the method of Turney and Littman (2003) for calculating the semantic orientation of a term defined by proximity to other terms in two polar dictionaries. |
| 7.embeddings/WordEmbeddings.ipynb | Explore word embeddings using Gensim |
| 8.neural/MLP.ipynb | MLP for text classification (keras) |
| 8.neural/ExploreMLP.ipynb | Explore MLP for your data (keras) |
| 8.neural/CNN.ipynb | CNN for text classification (keras) |
| 8.neural/LSTM.ipynb | LSTM for text classification (keras) |
| 8.neural/Attention.ipynb | Attention over word embeddings for document classification (keras) |
| 8.neural/AttentionLSTM.ipynb | Attention over LSTM output for text classification (keras) |
| 9.annotation/IAAMetrics.ipynb | Calculate inter-annotator agreement (Cohen's kappa, Krippendorff's alpha) |
| 10.wordnet/ExploreWordNet.ipynb | Explore WordNet synsets with a simple method for finding in a text all mentions of all hyponyms of a given node in the WordNet hierarchy (e.g., finding all buildings in a text). |
| 10.wordnet/Lesk.ipynb | Implement the Lesk algorithm for WSD using word embeddings |
| 10.wordnet/Retrofitting.ipynb | Explore retrofit word vectors |
| 11.pos/KeyphraseExtraction.ipynb | Keyphrase extraction with tf-idf and POS filtering |
| 11.pos/POS_tagging.ipynb | Understand the Penn Treebank POS tags through tagged texts |
| 12.ner/ExtractingSocialNetworks.ipynb | Extract social networks from literary texts |
| 12.ner/SequenceLabelingBiLSTM.ipynb | BiLSTM + sequence labeling for Twitter NER |
| 12.ner/ToponymResolution.ipynb | Extract place names from text, geolocate them and visualize on map |
| 13.mwe/JustesonKatz95.ipynb | Implement Justeson and Katz (1995) for identifying MWEs using POS tag patterns |
| 14.syntax/SyntacticRelations.ipynb | Explore dependency parsing by identifying the actions and objects that are characteristically associated with male and female characters. |
| 15.coref/CorefSetup.ipynb | Install neuralcoref for coreference resolution |
| 15.coref/ExtractTimeline.ipynb | Use coreference resolution for the task of timeline generation: for a given biography on Wikipedia, can you extract all of the events associated with the people mentioned and create one timeline for each person? |
| 16.ie/DependencyPatterns.ipynb | Measuring common dependency paths between two entities that hold a given relation to each other |
| 16.ie/EntityLinking.ipynb | Explore named entity disambiguation and entity linking to Wikipedia pages. |
| 17.clustering/TopicModeling_TODO.ipynb | Explore topic modeling to discover broad themes in a collection of movie summaries. |