multimodal social media content (text, image) classification

Implementation of an ETL process for real-time sentiment analysis of tweets with Docker, Apache Kafka, Spark Streaming, MongoDB and Delta Lake

Kaggle Twitter US Airline Sentiment, Implementation of a Tweet Text Sentiment Analysis Model, using custom trained Word Embeddings and LSTM-Deep learning [TUM-Data Analysis&ML summer 2021] @adrianbruenger @stefanrmmr

Tweet Text Writer Recognition Application

System developed by team datamafia in WNUT 2020 Task 2: Identification of informative COVID-19 English Tweets

Simple Repository regarding tweet classification using huggingface tokenizer and transformer, and tracking using weights and biases.

ML model to extract the main concept from a tweet trained on a dataset built with Babelnet and Babelfy.

Build a classifier that can extract a useful tweet among a ton of tweets during COVID-19 epidemic.

Treatment evaluation in presence of large number of covariates or treatment heterogeneity through Machine Learning methods

NLP Course By Deep learning.io powered by @coursera. Taught by: Younes Bensouda Mourri, Instructor of AI at Stanford University and Łukasz Kaiser, Staff Research Scientist at Google Brain.

This is a self-created basic tutorial for the Machine Learning II course at the Language Analysis and Processing master's degree where I show how I retrieve tweets from Twitter App and visualize, cluster and classify data by means of ML techniques and algorithms.

Using Pandas, Pickle, ReGex, Tweepy, Scikit-Learn, Sastrawi. NLTK, and bs4

EDA and Modeling Attempt for multi class text classification.

Welcome to our project, where we leverage advanced sentiment analysis techniques to detect and classify toxic content in game-related tweets. Our goal is to develop a predictive model that can accurately identify toxicity based on the language used in these tweets.

We aim to clasify tweets based on three categories 0: hate, 1: offensive, 2: neutral. For this purpose we use HateBert pretrained model with RNNs as the trainable layers

Classify whether a tweet constitutes a rumour event

This Jupyter Notebook demonstrates the implementation of a K-Nearest Neighbors (KNN) algorithm using the concept of nearest neighbors without using direct classifiers. It also includes exploratory data analysis (EDA) and comparison of three classifiers.

Exploring Jaccard-similarities technique on tweets then visualizing its output using k means and k means with pca. Additional input on time series analysis, web scrapping and twitter Scrapping.

Disaster Tweets API - Production-ready BERT-based FastAPI service for classifying tweets as Disaster or Not Disaster, with Docker, CI, and cloud deployment support.

PyTorch implementation of LSTM and Bi-LSTM networks for classifying personal health mentions in Twitter data