Data-Driven Road Safety: Accurate Traffic Accidents Hotspot Prediction
Using the latest Temporal Convolutional Networks on NHTSA data, we've set a new standard in traffic accident trend analysis, pinpointing risk areas with high accuracy in USA. Consequently, empowering transportation authorities to proactively mitigate accidents, surpassing conventional methodologies.
Leaf Area Index Prediction Using Temporal Deep Learning Models
A deep learning method to decode temporal patterns within a comprehensive array of landcover types. These models were applied to cloud-based datasets spanning a substantial 34-year timeframe to understanding of vegetation structure's implications on global climate dynamics by delivering accurate predictions of leaf area index.
Graph Convolutional Network classification on Cora dataset
Implemented a Graph Convolutional Network combining graph with a Neural Network to predict the topic of a publication given its citations using TensorFlow on CORA dataset.
Predicting Crop Yields in the United States by Integrating Deep Learning with GIS
A deep learning framework in Python that predicts crop yields across the U.S. with precision, drawing on over 16 years of remote sensing data and harnessing deep Gaussian Process.
Data Science in Healthcare - Disease Prediction based on patient's 24-hour ICU data
Evaluated the performance of ensemble methods to predict the patient’s health based on his/her 24-hour Intensive Care Unit data using Machine Learning algorithms.
Foliox - Conversational AI and Computer Vision Data Chatbot
A finance chatbot utilizing Conversational AI techniques to provide real-time stock market trends to users.