text-classification-multi-label

Repository for My HuggingFace Natural Language Processing Projects

Job title Classification by industry using NLP Multi-text Classification Problem.

Typed, calibrated decisions from any open LLM, read straight from its next-token probabilities.

IdiomX is a research-oriented Natural Language Processing (NLP) project that investigates the translation of English idiomatic expressions using transformer-based neural models.

Fine-tuned mBERT-based AI system for Urdu sentiment analysis with real-time Gradio UI and Hugging Face deployment.

Moodix — локальный модуль анализа русскоязычного текста, определяющий основное настроение (позитивное, нейтральное, негативное), 15 суб-настроений и 6 деструктивных признаков (угроза, ненависть, экстремизм и др.). Основан на BiLSTM-модели и работает без доступа к интернету. Подходит для интеграции в CRM, e-commerce, модерации и аналитики.

End-to-end NLP pipeline for categorising free-text medical claims using TF-IDF vectorisation and Random Forest classification.

Java + Python NLP pipeline: Selenium data collection, PostgreSQL, FastAPI classifier, BART-MNLI zero-shot classification

A machine learning-based sentiment analysis project that classifies text into Positive, Negative, and Neutral sentiments. The project uses TF-IDF vectorization and a trained classification model to generate sentiment predictions along with probability scores. Built as part of my AI/ML internship, with a focus on NLP preprocessing, model

This repository provides a machine learning and deep learning pipeline for text emotion detection. It includes a pretrained LSTM model, tokenizer, and preprocessing steps to classify emotions such as joy, sadness, and anger from text input. Easily deployable with provided resources and scripts.

Classify mental health text into multiple categories with fine-tuned BERT.

A BERT-based multi-label NLP system for detecting toxic and hateful language using the Jigsaw Toxic Comment dataset.

Clasificación de autoría de tweets mediante NLP, embeddings contextuales y Machine Learning, con MexSmall alcanzando 64.13% de accuracy.