NLP with Python: Glossary

Key Points

Language processing and Python	Undestanding Python and NLP
Text corpora and lexical resources	Accessing text corpora. Conditional Distribution Frequency (CDF) Re-use the Python codes. Lexical sources WordNet
Processing raw text	Lessons are design in four stages: conceptual, summative, formative, and connective.
Writing structured program	Lessons are design in four stages: conceptual, summative, formative, and connective.
Categorization and words tagging	Lessons are design in four stages: conceptual, summative, formative, and connective.
Learning to classifying text	Lessons are design in four stages: conceptual, summative, formative, and connective.
Extracting information from text	Lessons are design in four stages: conceptual, summative, formative, and connective.
Analyzing structure of sentence	Lessons are design in four stages: conceptual, summative, formative, and connective.
Building features based on grammar	Lessons are design in four stages: conceptual, summative, formative, and connective.
Analyzing the meaning of sentence	Lessons are design in four stages: conceptual, summative, formative, and connective.
Managing linguistic data	Lessons are design in four stages: conceptual, summative, formative, and connective.

Glossary

Annotated Corpus

Corpora to which some information is attached (POS, word boundaries, etc)

Arc

Connector of two nodes,represents a partial structure between them. Buffer of arcs to be added to the chart in future (stack)

Bag-of-words

Modeling a text (such as a sentence or a document) which is represented as the bag (multiset) of its words, disregarding grammar and even word order but keeping multiplicity.

Chart

Figure showing a process of parsing (nodes and arcs)

Corpus

A collection of (text) sentences, usually containing a large number of sentences.

Chunking

how to combine word into phrase/sentences

Corereference

Given a corpus, one refer another

Dependency annotated

annotation based on dependency in a sentence

Document frequency

Number of documents containing an index term j

Dictionary

Database of words including various kinds of information

Discourse (analysis)

Analyze not a single sentence but a set of sentences

Grammar

Set of rule/formalism that is used to specify what sentence are possible in language

Hash function, H(q)

Functino calculating the index for a search word q

History

List of arcs showing all generated arcs during parsing

Indexing

Procedure to extract index from documents

Information retrieval (IR)

process to find documents which are relevant to an users’s query from a document collection

Information (Compiled in Dictionary

Necessary information for NLP i.e: Morphological analysis, Syntatic analysis, Semantic analysis

Inverted Index

List of index term for each document in inverted form (row=index, column=document)

Lexicon

Dictionary of words and its type

Multimodal data

Data collected from various way from the same target, e.g: text, speech, video

Part-of-speech (POS) Tag

A word can be classified into one or more of a set of lexical or part-of-speech categories such as Nouns, Verbs, Adjectives and Articles, to name a few. A POS tag is a symbol representing such a lexical category - NN (Noun), VB (Verb), JJ (Adjective), AT (Article). One of the oldest and most commonly used tag sets is the Brown Corpus

Parsing

How to determine the structure of sentence according to grammar

Parse Tree

A tree defined over a given sentence that represents the syntactic structure of the sentence as defined by a formal grammar.etail below

Parallel Corpus

Collection of (translation) corpus of two or more language in parallel.

PCFG (Probabilistic Context Free Grammar)

Product (multiplication) of probabilities of rules in parse tree

Precision

Number of relevant documents in system output per number of documents in system output (how many selected items are relevant?)

Recall

Number of relevant document in system output per number of all relevant document in overall collection (how many relevant are the selected items?)

Relevance feedback

Asking user to judge whether retrieved documents are relevant or not

Sentence

An ordered sequence of tokens.

Semantic (role labelling)

Labelling sentence to give (different) meaning e.q: /I saw a man/ with telescope/ or /I saw/ a man with telescope/

Semantic primitive

Basic primitive of sense of word, e.g: human, abstract, concrete

Stop word

word not to be index term (function word; be, have, symbol; etc)

Syntactic

Relation of word to other words (structure).

Thesaurus

Database showing semantic relation between words

Tokenization

The process of splitting a sentence into its constituent tokens. For segmented languages such as English, the existence of whitespace makes tokenization relatively easier and uninteresting. However, for languages such as Chinese and Arabic, the task is more difficult since there are no explicit boundaries. Furthermore, almost all characters in such non-segmented languages can exist as one-character words by themselves but can also join together to form multi-character words.

Token

Before any real processing can be done on the input text, it needs to be segmented into linguistic units such as words, punctuation, numbers or alphanumerics. These units are known as tokens.

Trie

A tree structure such that common prefix of words are shares

N-gram

Collocation frequency of n words, e.g: 2-gram, 3-gram, etc.

N-gram model

Markov model for generation probability of a sentence by calculating probality of a word Ci depends only preceding N-1 words (for 2-gram)

Morphology (Morphological analysis)

Divide word into morphemes

Morphemes

Minimum linguistic Unit smaller than word, e.g. play-ing, un-kind-ly.

Named entity recognition

Given sentence, to identify the person, location, organization

Natural language generation

Generation of text from semantic interpretation

Node

Virtual node between words,

Vector space model

Vector represented both documents and query (from inverted index)

Query

An index terms or combination of index term

Query expansion

Automatic procedure to add related words to a query. E.g: Q=(car) –> Q=(car, automobile, auto, motorcar)

Zipf’s law

Law about distribution of frequency of English words. There are two ways to count words: type (how many types of word) and token (how many word apperas)