Music Genre Classification with Turkish Lyrics

Full code and most data (in accordance with CrowdTangle’s Terms of Service) supporting an article on what would remain on French-language Facebook if news content was removed

This project focuses the implementation on the Healthcare based chat-bot that answers the customer's queries as a therapist that is trained on the interaction between the patient and the therapist responses with the counsel chat dataset.

This GitHub repository hosts my AI evaluation work, featuring a Kaggle dataset analysis, experiments with three ML algorithms (including hyperparameter tuning), and a detailed exploration of wine quality data through outlier detection, correlation, and normalization techniques.

This repo contains everything about transformers and NLP.

Master Thesis Project - different phases of analyses developed for the LangLearn shared task (EVALITA 2023).

RAG architecture to retrieve and embed pdfs

Sentiment Analysis performed on Amazon reviews - polarity

Course recommendation system for students and learners willing to pursure similar courses. The provided url is on streamlit community cloud and is Live , uses bert embeddings and cosine smilarity

Implementing Multilingual WSD using [Normal, Atten]BiLSTM, Seq2Seq[Atten], Multitask WSD

CBERTdp is a strategy to speed up the clssification task by clustering BERT embeddings using different methods in order to use K-Means and the Dot-Product to obtaint the prediction results

example app for classification tasks

BERT-based senti- ment annotation is used to create unbiased datasets and then hybridize RNN with LSTM to find calculated ratings based on this unbiased reviews dataset.

This research proposes an ​​"AI-Driven Cross-Cultural Commodity Expert" framework​​, which addresses three critical technical bottlenecks through synergistic innovations in multilingual sentiment analysis, cultural quantification engines, and dynamic knowledge graphs

System for detecting fake news and suggesting credible alternatives. Takes a news URL and outputs a credibility score, explanation, and top reliable sources. Uses TF-IDF + Logistic Regression, XGBoost, and DistilBERT with hybrid BERT–LightGCN models, plus SHAP and GNNExplainer for interpretability.

This repository contains my scripts, results and visualization for my bachelor thesis "Medical concept PROBLEM: Polarity, Modality and Temporal Relations" - ON GOING, doing some code reorganize

Evaluating context-sensitive word meaning understanding in pair of sentences for BERT and GloVe+BiLSTM using WiC dataset | A3 for COL772 course (Fall 21)

Gender Classification problem - Predict if blog texts are written by male or female

Code, data and results of a reproducibility experiment for the paper "Exploring the Representation of Word Meanings in Context" by Marcos Garcia.

A Python-powered system for automatically categorizing and recommending Massive Open Online Courses using advanced NLP techniques and clustering algorithms.