Zaheeda Duduzile Chauke's Projects
Won 2nd Place: You are provided with a dataset of 500,000 reviews. Each review has a user provided rating from 1 to 5. The task is to build a model that can accurately predict the rating given only the text. As this is an NLP task you are permitted to use pretrained language models or word embeddings. You are not however permitted to use other sentiment datasets.
A robot powered training repository :robot:
Utilities for working with image data, text data, and sequence data.
Public repo for DeepLearning.AI MLEP Specialization
Financial Inclusion in Ghana: Using Machine Learning to bank the unbanked
Placed 13th out of 50 world-wide participants. The challenge of this competition will be to create a classifier that labels text content by the 27 SDG 3 indicators that are most “relevant.” The training data for this challenge includes approximately 3,000 web-scraped text from tenders, programs, and documents, as well as news articles about international development and humanitarian aid, and finally text descriptions of organizations involved in those sectors.