🏠 Home
Benchmark
📊 Tutti i benchmark 🦖 Dinosauro v1 🦖 Dinosauro v2 ✅ App To-Do List 🎨 Pagine libere creative 🎯 FSACB - Ultimate Showcase 🌍 Benchmark traduzione
Modelli
🏆 Top 10 modelli 🆓 Modelli gratuiti 📋 Tutti i modelli ⚙️ Kilo Code
Risorse
💬 Libreria di prompt 📖 Glossario IA 🔗 Link utili

Glossario IA

Il dizionario completo dell'Intelligenza Artificiale

162
categorie
2.032
sottocategorie
23.060
termini
📖
termini

Aspect-Based Sentiment Analysis (ABSA)

An NLP task that aims to identify specific aspects of an entity and determine the sentiment (positive, negative, neutral) expressed towards each of these aspects.

📖
termini

Aspect Category

A predefined classification of aspects into semantic groups (e.g., 'SERVICE', 'AMBIENCE', 'FOOD' for a restaurant) to structure the analysis.

📖
termini

Aspect Term

The explicit word or phrase in the text that refers to a specific aspect (e.g., 'the pizza', 'the service', 'the delivery time').

📖
termini

Sentiment Polarity

The classification of sentiment associated with an aspect into discrete categories, typically positive, negative, or neutral.

📖
termini

Target Opinion Detection

The task of linking an opinion or sentiment expression to the specific aspect (the target) it refers to in the text.

📖
termini

Dependency Parser

An NLP tool used to build the syntactic tree of a sentence, allowing descriptive words (adjectives) to be linked to their respective nouns (aspects).

📖
termini

Supervised Learning for ABSA

The use of machine learning models trained on manually annotated corpora to predict aspects and sentiment polarities.

📖
termini

Weakly Supervised Learning

Methods that use partial, noisy labels or heuristics to train ABSA models without requiring complete and costly manual annotation.

📖
termini

Sequence-to-Sequence (Seq2Seq) Models

A neural network architecture, often based on LSTM or Transformers, used to generate structured aspect-sentiment pairs from input text.

📖
termini

Convolutional Neural Networks (CNN) for ABSA

The application of CNNs to extract local features (n-grams) in text, effective for capturing relationships between nearby words like an adjective and its noun.

📖
termini

BERT for ABSA

The adaptation of the pre-trained language model BERT (Bidirectional Encoder Representations from Transformers) for ABSA tasks, often by adding a specific classification head.

📖
termini

Implicit Sentiment Analysis

The detection of sentiments that are not expressed by explicit polarity words but inferred from context (e.g., 'I couldn't eat my soup, it was cold' implies a negative sentiment about the 'temperature' aspect).

📖
termini

Annotated Corpus for ABSA

A set of text data where aspects, aspect terms, and sentiment polarities have been manually labeled, serving as a reference for training and evaluating models.

📖
termini

F1-Score Evaluation

A common evaluation metric for ABSA, which calculates the harmonic mean of precision and recall, measuring the model's ability to correctly identify aspect-sentiment pairs.

📖
termini

Implicit Aspect Detection

The identification of aspects that are not directly named but are implied by context (e.g., 'I waited 30 minutes' implies the 'waiting time' aspect).

📖
termini

Sentence-Level vs. Aspect-Level Sentiment Analysis

The distinction between classifying the overall sentiment of a sentence and the granular analysis of ABSA, which can reveal contradictory sentiments within the same sentence.

📖
termini

Quadruple Task in ABSA

An advanced formulation of ABSA that aims to simultaneously identify the opinion holder, aspect category, aspect term, and sentiment polarity in a text.

🔍

Nessun risultato trovato