Behavioral segmentation of open field in DeepLabCut, or B-SOID ("B-side"), is a pipeline that pairs unsupervised pattern recognition with supervised classification to achieve fast predictions of behaviors that are not predefined by users.

Filtering the stockmarket into clusters with financial ratios built on quarterly snapshots over 5 years of data. The goal was to see the evolution paths of companies over time and find clusters that develop future winners. Methods used, PCA, UMAP, HDBSCAN, KMeans.

Document Clustering, Summarisation and Visualisation on 20NewsGroup

Celem projektu było wykonanie analizy budżetów partycypacyjnych polskich miast pod kątem istniejących w nich wspólnych tematów, trendów, zależności.

Topic Modelling of News Dataset

Group project for the NUS module "IT1244 Artificial Intelligence: Technology and Impact"

Repository for fine-tuning and clustering sentence embeddings for Food Items

By utilizing various features, the Project assists the organization in categorizing countries and determining which country the NGO should allocate funds to.

Adapted BERTopic pipeline for Topic Modeling the arXiv dataset