Sentiment Analysis and NLP Concepts

Assigning a grammatical category (like noun, verb, adjective) to each word in a text.

A set of words that have a fixed membership and represent grammatical functions (e.g., prepositions, conjunctions).

A collection of pre-defined tags used to label words based on their grammatical function.

Using manually created linguistic rules to determine the grammatical category of each word in a sentence.

Words that can have multiple grammatical categories depending on the sentence context.

A probabilistic model that assumes the next word's category only depends on the current word's category.

Evaluating how well a POS tagging model performs by calculating measures like precision, recall, and F-measure.

The task of classifying each word in a sequence considering the relationship between neighboring words.

A variant of a Hidden Markov Model (HMM) where transitions are only allowed from a state to itself or to the next state; used in speech recognition.

A type of HMM where transitions are allowed between any two states, creating a fully connected network.

A probabilistic context-free grammar (PCFG) where each production rule has an associated probability, allowing for the prediction of different sentence structures.

A collection of sentences annotated with their corresponding parse trees, used for training and evaluating parsing models.

The probability of observing a specific output symbol (like a word) given a particular hidden state in a Hidden Markov Model.

An algorithm used in HMMs to calculate the probability of observing a specific sequence of output symbols.

An algorithm used in HMMs to determine the most likely sequence of hidden states that produced a given sequence of output symbols.

A dynamic programming structure used in the Forward and Viterbi algorithms to represent all possible states and transitions in an HMM.

A probabilistic model that assigns probabilities to parse trees, reflecting the likelihood of different syntactic structures.

In statistical parsing, the probability of a sentence is determined by summing up the probabilities of all its possible derivations.

A parsing approach that starts at the root symbol (S) of a parse tree and applies grammar rules to generate downward branching trees, exploring all possibilities with potential for generating invalid trees.

A parsing approach that begins with the terminal symbols of a sentence (words) and works upwards, applying grammar rules to combine them into phrases and eventually reach the root symbol (S). This approach avoids generating invalid trees but may fail to find complete parses.

The harmonic mean of precision and recall, providing a combined measure of how well a parsing model's results align with gold standard parse trees.

Opinion mining is an umbrella term for sentiment analysis, encompassing techniques for extracting and analyzing subjective information from text.

Sentiment analysis categorizes text into three levels: document level (overall sentiment), sentence level (individual sentence sentiment), and entity and aspect level (sentiment towards specific features or entities).

Sentiment analysis aims to classify text into three categories: positive, negative, or neutral, reflecting the emotional tone expressed.

The Bag of Words model simplifies text by representing it as a collection of unordered words, where the frequency of occurrence determines importance.

Negation words like 'not' or 'don't' present a challenge in sentiment analysis, as they can reverse the intended sentiment of a phrase.

The process of converting spoken language into written text using algorithms and technologies.

A method of identifying different speakers within a recording, determining who spoke at which time.

Analyzing speech to detect emotions like happiness, anger, sadness, etc.

The use of artificial intelligence to convert text into natural-sounding speech.

A Python library that provides an interface to the Google Web Speech API, allowing you to perform speech recognition.

An open-source speech recognition engine that works offline, perfect for situations without internet access.

Audio formats commonly used for speech recognition tasks.

The number of samples taken per second to capture sound, influencing audio fidelity and file size.

Sentiment analysis often uses social media posts, product reviews, and news articles to gather data about public opinion.

Pointwise Mutual Information (PMI) measures how often two words appear together compared to their individual frequencies, indicating word associations or collocations.

NLTK is a Python library widely used for working with natural language, including sentiment analysis.

Named Entity Recognition (NER) in NLP categorizes entities like people, places, and organizations from text, making sense of who, what, and where.

Dictionary-based NER uses predefined lists of words or phrases to match entities in text.

Machine learning-based NER models learn from data to recognize entities, using statistical techniques and features.

Deep learning-based NER utilizes complex neural networks to extract entities, capturing intricate relationships in text.

Chunk analysis groups words into meaningful phrases like noun phrases, using IOB tagging to mark the beginning, inside, or outside of a chunk.