Computational Linguistics - Intermediate Course

Explore lemmatization, reducing words to their base forms to unify morphology, using dictionary-based and rule-based methods with examples like runs and running becoming run, and cats becoming cat.

Stemming trims word endings to a base form, enabling faster information retrieval, uses stemmers like Port, Snowball, and Lancaster, and ignores context or dictionary checks.

Learn how to assign parts of speech to each word in a text, using rule-based, statistical, and machine learning taggers, and explore challenges like context, out-of-vocabulary words, and multilingual texts.

Learn named entity recognition, the NLP process that identifies and classifies entities like persons, organizations, locations, dates, monetary values, and percentages in text, supporting chatbots and information retrieval.

Explore sentiment analysis and opinion mining to identify emotional tone in words and sentences. Apply polarity and subjectivity insights using tools like TextBlob for real-world reviews and customer feedback.

Explore sentiment analysis, also called opinion mining, to detect emotional tone and polarity in text. See how context and domain specificity affect results, with TextBlob demonstrations.

Explore the NLTK Python library for natural language processing, covering tokenization, parsing, classification, stemming, tagging, semantic reasoning, and sentiment analysis, with corpora and WordNet resources.

Learn to prepare text for natural language processing by installing nltk, fetching text via the Wikipedia application programming interface in Google Colab, and preparing for tokenization and nlp workflows.

Explore tokenization with NLTK, covering word and sentence tokenization, importing and downloading the toolkit, tokenizing text into words and sentences, and handling punctuation in practical examples.

Learn lemmatization with NLTK by using the WordNet Lemmatizer, tokenizing sentences and lemmatizing words, and compare with stemming as introduced for upcoming videos.

Explore stemming in Python using the NLTK library by comparing porter and lancaster stemmers, tokenizing sentences, and observing how each stemmer yields different word reductions.

Learn to perform part-of-speech tagging with the NLTK library by tokenizing text, tagging tokens, and interpreting tag types for linguistic analysis.

Explore named entity recognition with NLTK, download required resources, tag tokens, and compare deterministic model outputs and common errors.

Remove stopwords with NLTK by importing Stopwords, tokenizing text, and filtering English words to reduce noise for effective NLP and subsequent sentiment analysis.

Explore sentiment analysis with nltk's vader sentiment intensity analyzer, download the vader lexicon, and build a function to score text with compound, positive, negative, and neutral results.

SpaCy is an open-source, production-ready, fast NLP library trained on large data from news and magazines, offering tokenization, POS tagging, named entity recognition, and parsing for real-world applications.

Learn how to tokenize text with spaCy by preparing a text sample, loading en_core_web_sm, creating a doc, and extracting word and sentence tokens.

Apply lemmatization with spaCy to reduce the first sentence's words to their lemmas using a doc and nlp object, then print the lemmas. Prepare to dive into POS tagging next.

Explore named entity recognition with spaCy, compare its performance to NLTK, and extract entities with labels such as person, cardinals, organizations, and money.

Explore sentiment analysis with spaCy and TextBlob, install and configure the spaCy TextBlob pipeline, analyze polarity and subjectivity across three sentences, and troubleshoot common errors.

Explore the Python Re library for regular expressions, using match, search, findall, sub, and split to search, replace, and extract with pattern-based rules; see examples for proper names and ages.

Learn how to structure regular expressions, using raw strings with an r prefix, escape characters, and core elements like identifiers, metacharacters, quantifiers, and character sets to match and extract data.

Learn how the r prefix defines raw strings in Python by placing R before the opening quote, ensuring literal interpretation for regular expressions and avoiding escape characters.

Explore how the Python 3 re module uses finditer, findall, match, and search to extract information from strings, with compiled patterns, raw strings, and iterators that return match objects.

Learn how to work with match objects from regular expressions, using compile, finditer, and group to extract span, start, end, and matched text.

Practice locating all occurrences of red in the given text using python's re module, compiling a pattern and using finditer with start and end to print positions.

Master using the re module to search a string for the word read, compile a regex, and print the start and end positions of each match with finditer.

Learn how identifiers in regular expressions represent character types, such as digits and whitespace, via extended backslash escapes and boundaries, as shown in a Python re example.

Extract all digits from the given string using the finditer function, store the resulting match objects in a variable, and print each with a for loop.

Solve exercise two by using Python's re module to find digits in a string, compiling a digit pattern and iterating over matches to print each start and end index.

Explore metacharacters in regular expressions, including the dot, caret, dollar sign, and vertical bar, and see how they define patterns and search at the beginning of a string.

Practice regular expressions by building a pattern with a metacharacter and an identifier to test if a string starts with a non-digit, and print the match object if so.

Demonstrates how to use a regular expression pattern with a metacharacter to locate a non-digit at the beginning of an input string, using re.compile, finditer, and printing any matches.

Demonstrate string parsing by locating all periods in a given text and printing each period position, using a regex approach on the sample string about John, Kathy, Peter, and Melissa.

Solve the exercise by compiling a pattern with an escaped period to match literal periods in the input text, iterate through matches, and print each start index with an f-string.

Explore quantifiers in regular expressions, including optional, star, plus, and braces for exact or range counts. Demonstrate using Python's re module with compile and finditer to identify three-or-more-character words.

Apply regex quantifiers to locate all numbers in a string, print every match object for exercise 5, and anticipate the next video with a detailed solution and explanation.

Find all numbers in a string using the regular expression \d+ and finditer to print match objects, including their groups and spans.

Learn how sets in regular expressions define character groups with square brackets, including vowels and inverted caret sets, and use ranges like A-C with one or more occurrences.

Learn how to modify strings in Python using split and sub, apply regular expressions with the re module, and practice splitting and replacing patterns to manipulate text.

Practice substitution and split techniques by replacing the name John with a female name, changing his to her, and splitting the text by numbers to form a list.

Create a pattern for John, replace John with Rebecca, and replace his with her; compile patterns, match digits, and split on the digit pattern to extract relevant parts.

Construct a dictionary of unique name-to-age key-value pairs from the text, using a name-number pattern, with entries like John 39, Michael six, Peter 104, and Melissa 98.

Compile a digits pattern D13 representing 1 to 3 digits and a names pattern A-Za-z+ using the module, use findall to extract ages and names and build a name-age dictionary.

Extract http:// URLs from a text using a regex, identify which among five sample URLs starts with http://, and print each match with the group method using the re module.

Identify and extract the urls that start with http skeptical smiley face using the Re library and regular expressions, including using raw strings, compiled patterns, and iterating over matches.

Explore NLP with generative AI by building text processing tasks using OpenAI APIs, including lemmatization, pos tagging, ner, and sentiment analysis, and compare with Spacey in a final project.

Download the kaggle dataset, load negative and positive txt files in Google Colab, and convert them into two Python lists; then generate and sample code for OpenAI API prep.

Learn how to obtain and securely handle an OpenAI API key to power your code with ChatGPT-style models, including creating, copying, and deleting the key.

Verify API access by reviewing the OpenAI text generation documentation, set up a Colab workflow to run completions, and experiment with temperature and end parameters for NLP tasks like lemmatization.

Develop a lemmatization function with generative AI, crafting a system prompt to guide the lemmatizer, then compare results with the original text and note stopword removal.

Perform pos tagging with generative ai by adapting a step four workflow, updating prompts and function names, and evaluating results against spacy.

Extract named entities using a generative AI workflow, modifying prompts and function names, processing multiple sentences, and comparing outputs with Spacey for named entity recognition.

Learn sentiment analysis with generative ai by crafting prompts that classify text as positive, negative, or neutral and output a source sentence with judgment.

Generate tailored responses to hotel reviews using sentiment analysis and a review response prompt, delivering negative sentiment with apologies, neutral feedback with gratitude, or positive praise.

Explore WordNet within NLTK as a rich lexical database of English words, detailing semantic relationships including hyponyms, hypernyms, meronymy, and antonymy, with similarity metrics and pos tagging filters.

Learn the main methods in WordNet, including synsets, definitions, and examples, plus hypernyms and hyponyms. Compare semantic similarity via Wu-Palmer and path similarity within the WordNet hierarchy.

Explore how to analyze a sentence with WordNet via NLTK, tokenize input, retrieve synsets, definitions, examples, and hypernyms or hyponyms, and measure Wu-Palmer and path similarity.

Explore advanced WordNet functionalities, including lemma names and lemmas, parts of speech, antonyms, and synset access by pos and offset, plus language support through langs.

Practice identifying fruit tokens with WordNet in NLTK by tokenizing an input string and filtering tokens that refer to fruits, yielding a list like bananas, oranges, apricots, avocados, watermelon.

solve an exercise by tokenizing the input string, using a synset fruit concept and WordNet hypernyms to identify fruit tokens, collect unique results, and print them.

Conclude by summarizing core NLP concepts and tools like NLTK, Spacy, regular expressions, and Jupyter notebooks in Anaconda, and emphasize applying these libraries to consulting projects in hotels and airlines.